Full text of "dec :: pdp8 :: handbooks :: IntroToProgramming1969"

See other formats

INTRODUCTION

TO
PROGRAMMING

PDP-8 Family Computers

Prepared by

The Software Writing Group

Programming Department

Digital Equipment Corporation

SMALL COMPUTER HANDBOOK SERIES

First Printing, January 1969
Second Printing, July 1969

PDP is a registered trademark
of Digital Equipment Corporation.

Foreword

The data processing industry has expanded so rapidly during the
past twenty years that there has always been a severe shortage of
trained* personnel. Many new job openings are created each year be-
cause new ways for using the computer are continually being developed.
Consequently, the computer industry has a real problem training pro-
grammers and engineers fast enough to fill these new jobs.

Fortunately, this situation is improving, primarily because our uni-
versities, high schools and training schools have greatly expanded their
abilities and facilities to train students in data processing. We are espe-
cially heartened by the great strides being made at the secondary school
level. Computers are now an integral part of modern life; high school
students are using computers to learn to solve Algebra I problems;
engineers are using the computer as an "electronic slide rule"; chemists
are using the computer to aid in analysis and control of chemical pro-
cesses; many businesses are using computers to process payroll records,
control inventories, and many other applications.

We anticipate that this book will be useful to both teachers and
students as a training text and reference handbook. For users of our
small computers, it will also be the basic programming reference source
for use in conjunction with the many small computer systems of the
PDP-8 family produced by DEC.

Introduction to Programming was prepared by the Software Writing
Group in the Programming Department at DEC, with the assistance
of many others, including instructors in our Training Department, many
DEC and user programmers who have reviewed the manuscript, and
teachers who are presently using a PDP-8 in their computer sciences
courses.

We are most grateful to everyone who has contributed.

Kenneth H. Olsen

President

Digital Equipment Corp.

Preface

This textbook is an introduction to programming digital computers,
particularly Digital Equipment Corporation's PDP-8 family of com-
puters (often referred to simply as PDP-8 1 ). Over 3,000 of these small,
general-purpose computers have proven their versatility in hundreds
of different applications.

Introduction to Programming can be used by students, programming
trainees, and experienced programmers. It offers two approaches to
learning computer programming: (1) learning to program in machine
language, that is, learning to write programs using the actual instruc-
tions that the computer was built to perform and (2) learning to pro-
gram in a common programming language that uses many English
words and standard mathematical notation. Although easier to learn,
common languages, such as FORTRAN, ALGOL or FOCAL, are not
as efficient as machine languages, in terms of program execution times
and core memory requirements. Of course, machine language and com-
mon language programming are complementary, so it is useful to learn
both.

Teachers and programming students will find Chapters 1 through 5
useful as an introductory text to machine language programming.
Chapters 6 through 9 are devoted to the descriptions, uses, and oper-
ating procedures for PDP-8 system software; this software has proven
its ability to simplify the tasks of writing, editing, assembling, compil-
ing, debugging, and running user programs.

Chapter 6 describes PDP-8 system software and provides detailed
operating procedures for the Symbolic Editor, the assemblers, and other
commonly used software. These operating procedures can be used by
the computer operator or programmer independent of the remainder
of the book.

Chapters 7 and 8 describe the Disk Monitor System and the TSS/8
Time-Sharing System.

1 PDP stands for Programmed Data Processors and is a trade mark of Digital
Equipment Corporation.

Chapter 9 is a complete student's text on the use of FOCAL
(FOrmula CALculator), a conversational interpreter for solving nu-
merical problems. FOCAL language consists of short, easy-to-learn,
imperative English statements. FOCAL puts the full calculating power
and speed of the computer at the user's fingertips, providing an easy
way of simulating mathematical models, plotting curves, handling sets
of simultaneous equations in n-dimensional arrays, and much more.

Scientific programming is explained in Chapter 10 along with a de-
tailed special program designed to gather physiological data.

After becoming familiar with PDP-8 programming, the user may
wish to join DECUS (Digital Equipment Computer Users Society).
DECUS is a user's organization that exchanges ideas and programs;
it is described in Chapter 1 1 together with a list of programs available
from DECUS.

A detailed index /glossary, a summary of instructions, answers to
selected exercises, and taWes of conversion codes are included at the
back of the book. More experienced programmers will find the book
and its index useful as a reference guide.

Preface to the Second Printing

The text is unchanged, except for the correction of minor errors.
Most of these errors were reported by diligent readers, to whom we
are most grateful. As we expect to improve this book in future re-
visions, all readers are earnestly requested to send corrections and
comments to:

Manager, Software Documentation
Programming Department
Digital Equipment Corporation
Maynard, Mass. 01754

Contents

Foreword . iii

Preface , v

Contents vii

Chapter 1 Computer Fundamentals 1-1

Introduction to PDP-8 Family Computers, Ap-
plications, Computer Number Systems and Arith-
metic and Logical Operations

Chapter 2 Programming Fundamentals 2-1

Memory Reference Instructions and Operate Mi-
croinstructions and the Way They Are Used in
Programming the PDP-8 Family Computers

Chapter 3 Elementary Programming Techniques 3-1

Phases of Program Preparation, Programming
Symbols and Conventions, Arithmetic Opera-
tions, Subroutines, Address Modification and
Auto-indexing, Program Looping and Branching

Chapter 4 System Description and Operation 4-1

Entering and Storing Information in Core Mem-,
ory with the Operator's Console and Teletype
Unit, followed by an Introduction to the More
Common Optional Equipment of the PDP-8
Family

Chapter 5 Input/Output Programming 5-1

I/O Transfer Instructions and I/O Programming
Techniques applied to the Teletype Unit Using the
ASCII Character Set, followed by Descriptions of
the Program Interrupt Facility and the Data Break
Option

Chapter 6 Operating the System Software 6-1

Descriptions of PDP-8 Family System Software
supplied by Digital Equipment Corporation and
Operating Procedures for the Symbolic Editor,
Assemblers, and Other Commonly Used Software

vii

Chapter 7 Disk Monitor System 7-1

Description of the Disk Monitor System, which
includes a Keyboard-Oriented Monitor, a FOR-
TRAN Compiler, an Editor, a Peripheral Inter-
change Program, and a Dynamic Debugging
Program

Chapter 8 Time-Sharing System 8-1

Description of the TSS/8 Time-Sharing System,
which includes a Monitor and a Comprehensive
Library with Facilities for Compiling, -Assembling,
Editing, Loading, Calling, and Debugging User
Programs

Chapter 9 FOCAL Programming 9-1

Complete student's text on the Use of FOCAL
(FOrmula CALculator), a Conversational Inter-
preter for Solving Numerical Problems, including
several Examples and Methods for Obtaining Spe-
cific Problem Solutions

Chapter 10 PDP-8 Family Computers in the Sciences 10-1

Discussion of several general Scientific Applica-
tions Using PDP-8 Family Computers, followed by
a Detailed Description of a Special Program De-
signed to Gather Physiological Data

Chapter 11 Digital Equipment Computer Users Society 11-1

Description of the Objectives and Functions of
DECUS, including the DECUS Program Library
and Catalog, DECUSCOPE, Activities, Member-
ship, and Policies and Administration

Appendix A Answers to Selected Exercises A-l

Appendix B Character Codes B-l

Appendix C Flowchart Guide C-l

Appendix D Tables of Instructions D-l

Appendix E Legal Microinstruction Combinations E-l

Appendix F Miscellaneous Tables F-l

Index/Glossary Index -1

Chapter I

Computer
Fundamentals

INTRODUCTION

During the past 20 years, the computer revolution has dramatically
cfianged our world, and it promises to bring about even greater changes
in the years ahead.

The general purpose, digital computers being built today are much
faster, smaller and more reliable and can be produced at lower cost
than the earlier computers. But even more significant breakthroughs
have come in the many new ways we have learned to use computers.

The first big electronic computers were usually employed as super
calculators to solve complex mathematical problems that had been im-
possible to attack before. In recent years, computer programmers
have begun using computers for non-numerical operations, such as
control systems, communications, and data handling and processing.
In these operations, the computer system processes vast quantities of
data at high speed.

The Computer Challenge

It has been said that a computer can be programmed to do any
problem that can be defined. The key word here is defined, which
means that the solution of the problem can be broken down into a
series of steps that can be written as a sequence of computer instruc-
tions. The definition of some problems, such as the translation of natu-
ral languages, has turned out to be very difficult. A few years ago it
was thought that computer programs could be written to translate
French into English, for example. As a matter of fact, it is quite easy
to translate a list of French words into English words with similar
meanings. However, it is very difficult to precisely translate sentences
because of the many shades of meanings associated with individual
words and word combinations. For this reason,- it is not practical to

1-1

try to communicate with a computer using a conventional spoken lan-
guage.

While natural languages are impractical for computer use, program-
ming languages, such as FOCAL, ALGOL, and FORTRAN with their
precisely defined structure and syntax, greatly simplify communication
with a computer. Programming languages are problem oriented and
contain familiar words and expressions; thus, by using a programming
language, it is possible to learn to write programs after a relatively
short training period. Since most computer manufacturers have adopted
standard programming languages and implemented the use of these
languages on their computers, a given program can be executed on a
large number of computers. PDP-8 programmers use FORTRAN and
ALGOL-8 for scientific and engineering problems and use FOCAL-8
and BASIC-8 for shorter numerical calculations. Computer languages
have been developed for programmed control of machine tools, com-
puter typesetting, music composition, data acquisition, and many other
applications. It is likely that there will be many more new programming
languages in the future. Each new language development will enable
the user to more easily apply the power of the computer to his partic-
ular problem or task.

Who can be a programmer? In the early days of computer program-
ming, most programmers were mathematicians. However, as this text
illustrates, most programming requires only an elementary ability to
handle arithmetic and logical operations. Perhaps the most basic re-
quirement for programming is the ability to reason logically.

The rapid expansion of the' computer field in the last decade has
made the resources of the computer available to hundreds of thousands
of people and has provided many new career opportunities.

Computer Applications

A computer, like any other machine, is used because it does certain
tasks better and more efficiently than humans. Specifically, it can re-
ceive more information and process it faster than a human. Often,
this speed means that weeks or months of pencil and paper work can
be replaced by a method requiring only minutes of computer time.
Therefore, computers are used when the time saved by using a com-
puter offsets its cost. Further, because of its capacity to handle large
volumes of data in a very short time, a computer may be the only
means of resolving problems where time is of the essence. Because of
the advantages of high speed and high capacity, computers are being
used more and more in business, industry, and research. Most com-
puter applications can be classified as either business uses, which usually

1-2

rely upon the computer's capacity to store and quickly retrieve large
amounts of information, or scientific uses, which require accuracy
and speed in performing mathematical calculations. Both of these are
performed on general purpose computers. Some examples of computer
applications are given below.

Solving Design Problems. The computer is a very useful calculating
tool for the design engineer. The wing design of a supersonic aircraft,
for example, depends upon many factors. The designer describes each
of these factors in the form of mathematical equations in a program-
ming language. The computer can then be used to solve these equations.

Scientific and Laboratory Experiments. In scientific and laboratory
experiments, computers are used to evaluate and store information
from numerous types of electronic sensing devices. Computers are par-
ticularly useful in such systems as telemetry where signals must be
quickly recorded or they are lost. These applications require rapid and
accurate processing for both fixed conditions and dynamic situations.

Automatic Processes. The computer is a useful tool for manufac-
turing and inspecting products automatically. A computer may be pro-
grammed to run and control milling machines, turret lathes, and many
other machine tools with more rapid and accurate response than is
humanly possible. It can be programmed to inspect a part as it is being
made and adjust the machine tool as needed. If an incoming part is de-
fective, the computer may be programmed to reject it and start the
next part.

Training by Simulation. It is often expensive, dangerous and imprac-
tical to train a large group of men under actual conditions to fly a
commercial airplane, control a satellite, or operate a space vehicle. A
computer can simulate all of these conditions for a trainee, respond to
his actions, and report the results of the training. The trainee can there-
fore receive many hours of on-the-job training without risk to himself,
others, or the expensive equipment involved.

Applications, such as those given above and in Chapter 10, often
require the processing of both analog and digital information. Analog
information consists of continuous physical quantities that can be easily
generated and controlled, such as electrical voltages or shaft rotations.
Digital information, however, consists of discrete numerical values,
which represent the variables of a problem. Normally, analog values
are converted to equivalent digital values for arithmetic calculations
to solve problems. Some computers, such as the LINC-8, combine the
analog and digital characteristics in one computer system.

1-3

Computer Capabilities and Limitations

A computer is a machine and, as all machines, it must be directed
and controlled in order to perform a useful task. Until a program is
prepared and stored in the computer's core memory, the computer
"knows" absolutely nothing, not even how to receive input. Thus, no
matter how good a particular computer may be, it must be "told"
what to do. The usefulness of a computer therefore can not be fully
realized until the capabilities (and the limitations) of the computer
are recognized.

Repetitive operation — A computer can perform similar operations
thousands of times, without becoming bored, tired or careless.
Speed — A computer processes information at enormous speeds,
which are directly related to the ingenuity of the designer and the
programmer. Modern computers can solve problems millions of
times faster then a skilled mathematician.

Flexibility — General purpose computers may be programmed to
solve many types of problems.

Accuracy — Computers may be programmed to calculate answers
with a desired level of accuracy as specified by the programmer.
Intuition — A computer has no intuition. It can only proceed as it
is directed. A man may suddenly find the answers to a problem
without working-out the details, but a computer must proceed as
ordered.

The remainder of this chapter is devoted to the general organization
of the computer and the manner in which it handles data. Included are
the number systems used in programming together with the arithmetic
and logical operations of the computer. This information provides a
necessary background for all who desire a basic appreciation of com-
puters and their uses, and it is a prerequisite to machine-language
programming, covered in chapters 2 through 5.

1-4

NUMBER SYSTEM PRIMER

The concept of writing numbers, counting, and performing the basic
operations of addition, subtraction, multiplication, and division has
been directly developed by man. Every person is introduced to these
concepts during his formal education. One of the most important
factors in scientific development was the invention of the decimal
numbering system. The system of counting in units of tens probably
developed because man has ten fingers. The use of the number 10 as
the base of our number system is not of prime importance; any stand-
ard unit would do as well. The main use of a number system in early
times was measuring quantities and keeping records, not performing
mathematical calculations. As the sciences developed, old numbering
systems became more and more outdated. The lack of an adequate
numerical system greatly hampered the scientific development of early
civilizations.

Two basic concepts simplified the operations needed to manipulate
numbers; the concept of position, and the numeral zero. The concept
of position consists of assigning to a number a value which depends
both on the symbol and on its position in the whole number. For
example, the digit 5 has a different value in each of the three numbers
135, 152, and 504. In the first number, the digit 5 has its original
value 5; in the second, it has the value of 50; and in the last number,
it has the value of 500, or 5 times 10 times 10. Sometimes a position
in a number does not have a value between 1 and 9. If this position
were simply left out, there would be no difference in notation between
709 and 79. This is where the numeral zero fills the gap. In the number
709, there are 7 hundreds, tens and 9 units. Thus, by using the
concept of position and the numeral 0, arithmetic becomes quite easy.

A few basic definitions are needed before proceeding to see how
these concepts apply to digital computers.

Unit — The standard utilized in counting separate items is the unit.

Quantity — The absolute or physical amount of units.

Number — A number is a symbol used to represent a quantity.

Number System — A number system is a means of representing
quantities using a set of numbers. All modern number systems use
the zero to indicate no units, and other symbols to indicate quan-
tities. The base or radix of a number system is the number of sym-
bols it contains, including zero. For example the decimal number
system is base or radix 10, because it contains 10 different sym-
bols (viz., 0,1,2,3,4,5,6,7,8, and 9).

1-5

Binary Number System

The fundamental requirement of a computer is the ability to physi-
cally represent numbers and to perform operations on the numbers
thus represented. Although computers which are based on other num-
ber systems have been built, modern digital computers are all based
on the binary (base 2) system. To represent ten different numbers
(0,1,2, . . . , 9) the computer must possess ten different states with
which to associate a digit value. However, most physical quantities have
only two states: a light bulb is on or off; switches are on or off; holes
in paper tape or cards are punched or not punched; current is positive
or negative; material is magnetized or demagnetized; etc. Because it
can be represented by only two such physical states, the binary number
system is used in computers.

To understand the binary number system upon which the digital
computer operates, an analysis of the concepts underlying the decimal
number system is beneficial.

POSITION COEFFICIENT

In the decimal numbering system (base 10), the value of a numeral
depends upon the numeral's position in a number, for example:

347 = 3 X 100 = 300
4 X 10 = 40

7 X 1 = 7

347

The value of each position in a number is known as its position coeffi-
cient. It is also called the digit position weighting value, weighting value,
or weight, for short. A sample decimal weighting table follows:

10 s 10 2 10 1 10°

and, as shown above,

347 = 3 X 10 2 4- 4 X 10 1 + 7 X 10°.

Weighting tables appear to serve no useful purpose in our familar deci-
mal numbering system, but their purpose becomes apparent when we
consider the binary or base 2 numbering system. In binary we have
only two digits, and 1. In order to represent the numbers 1 to 10, we
must utilize a count-and-carry principle familar to us from the decimal

1-6

system (so familiar we are not always aware that we use it). To count
from to 10 in decimal, we count as follows:

2
3
4
5
6
7

10 with a carry to the 10 1 column

Continuing the counting, when we reach in the units column again,
we carry another 1 to the tens column. This process is continued until
the tens column becomes and a 1 is carried into the hundreds column,
as shown below:

•

.0 one carry

one carry

100 two carries

COUNTING IN BINARY NUMBERS

In the binary number system, the carry principle is used with only
two digit symbols, namely and 1. Thus, the numbers used in the
binary number system to count up to a decimal value of 10 are the
following.

Binary

Decimal

Binary

Decimal

(0)

110

(6)

(1)

111

(7)

(2)

1000

(8)

(3)

1001

(9)

100

(4)

1010

(10)

101

(5)

When using more than one number system, it is customary to subscript
numbers with the applicable base (e.g., 101 2 =5 10 ).

1-7

A weighting table is used to convert binary numbers to the more
familiar decimal system.

2 4 2» 22 2 1 2° (Weight Table)
10 10 1 (Binary Number)
Digit

* = 1

X
X
X
X
X

Position
Coefficient

Decimal Number =21

It should be obvious that the binary weighting table can be extended,
like the decimal table, as far as desired. In general, to find the value
of a binary number, multiply each digit by its position coefficient and
then add all of the products.

ARRANGEMENTS OF VALUES

By convention, weighting values are always arranged in the same
manner; the highest on the extreme left and the lowest on the extreme
right. Therefore, the position coefficient begins at 1 and increases from
right to left. This convention has two very practical advantages. The
first advantage is that it allows the elimination of the weighting table,
as such. It is not necessary to label each binary number with weighting
values, as the digit on the extreme right is always multiplied by 1, the
digit to its left is always multiplied by 2, the next by 4, etc. The second
advantage is the elimination of some of the 0s. Whether a is to the
right or left, it will never add to the value of the binary number. Some
0s are required, however, as any 0s to the right of the highest valued
1 are utilized as spaces or place keepers, to keep the Is in their correct
positions. The 0s to the left, however, provide no information about
the number and may be discarded, thus the number 0001010111 =
1010111.

The PDP-8 family computers operate upon 12-bit (binary digit)
numbers. This means that the numbers from to 11111111111 1 2
(4095 10 ) can be directly represented.

SIGNIFICANT DIGITS

The "leftmost" 1 in a binary number is called the most significant
digit. This is abbreviated MSD. It is called the "most significant" in
that it is multiplied by the highest position coefficient. The least sig- -
nificant digit, or LSD, is the extreme right digit. It may be a 1 or 0,
and has the lowest weighting value, namely 1. The terms LSD and

1-8

MSD have the same meaning in the decimal system as in the binary sys-
tem, as shown below.

^_J^1 110 10 1,

MSD .<^7o 10 10 < l^> LSD

~~-»4 5, 9 7 1

CONVERSION OF DECIMAL TO BINARY

There are two commonly used methods for converting decimal num-
bers to binary equivalents. The reader may choose whichever method
he finds easier to use.

1. Subtraction of Powers Method — To convert any decimal number
to its binary equivalent by the subtraction of powers method, proceed
as follows.

Subtract the highest possible power of two from the decimal number,
and place a "1" in the appropriate weighting position of the partially
completed binary number. Continue this procedure until the decimal
number is reduced to 0. If, after the first subtraction, the next lower
power of 2 cannot be subtracted, place a in the appropriate weight-
ing position. Example:

£io

= ? binary

- 8

-2

2 3

2 2

2°

Power

Value

Binary

Therefore, 42 10 = 101010,.

2. Division Method — To convert a decimal number to binary by
the division method, proceed as follows.

Divide the decimal number by 2. If there is a remainder, put a 1 in the
LSD of the partially formed binary number; if there is no remainder,
put a in the LSD of the binary number. Divide the quotient from the
first division by 2, and repeat the process. If there is a remainder,

1-9

record a I; if there is no remainder, record a 0. Continue until the
quotient has been reduced to 0. Example:

47 10 = ? Binary

5TT
7T
TT
TT

Quotient

23
11

Remainder

1 —
—

V V V V

1111

Therefore, 47„ = 101111,.

EXERCISES

a. Decimal-to-Binary Conversion — Convert the following decimal
numbers to their binary equivalents.

15*

11.

4095

18.

12.

1502

42 tt

13.

377„

100„

14.

501*,

235,„

15.

828 10

lxo

16.

907 10

294 10

17.

4000,

117„

18.

3456,

86 IO

19.

2278,

10.

4090 10

20.

1967,

b. "Binary to Decimal Conversion
numbers to their decimal equivalents.

Convert the following binary

110,

11011011101,

101,

10.

111000111001,

1110110,

11.

111010110100,

1011110,

12.

111111110111,

0110110.

13.

101011010101,

11111 2

14.

111111,

1010 2

15.

000101001 2

110111,

16.

111111111111,

1-10

Octal Number System

It is probably quite evident at this time that the binary number
system, although quite nice for computers, is a little cumbersome for
human usage. It is very easy for humans to make errors in reading and
writing quantities of large binary numbers. The octal or base 8 num-
bering system helps to alleviate this problem. The base 8 or octal num-
ber system utilizes the digits through 7 in forming numbers. The
count-and-carry method mentioned earlier applies here also. Table 1-1
shows the octal numbers with their decimal and binary equivalents.

Table 1-1. Decimal-Octal-Binary Equivalents

Decimal

Octal

Binary

Decimal

Octal

Binary

111

1000

1001

1010

100

1011

101

1100

110

1101

The octal number system eliminates many of the problems involved
in handling the binary number system used by a computer. To make the
12-bit numbers of the PDP-8 computers easier to handle, they are
often separated into four 3-bit groups. These 3-bit groups can be rep-
resented by one octal digit using the previous table of equivalents as
seen below.

A binary number 11010111101

is separated into 3-bit groups by starting with the LSD end of the

number and supplying leading zeros if necessary:
011 010 111 101

The binary groups are then replaced by their octal equivalents:

011 2 = 3 e

010 2 = 2 8

111,= 7 8

101,= 5 8

and the binary number is converted to its octal equivalent:

3 2 7 5.
Conversely, an octal number can be expanded to a binary num-
ber using the same table of equivalents.

5307 8 = 101 011 000 'ill,

1-11

OCTAL-TO-DECIMAL CONVERSION

Octal numbers may be converted to decimal by multiplying each
digit by its weight or position coefficient and then adding the resulting
products. The position coefficients in this case are powers of 8, which
is the base of the octal number system. Example:

2167,=

? decimal

2167 8 =

7 X 8° = 7 X

1 =

+6 X 8i = 6 X

8 =

+ [ X 8 2 = 1 X

64 =

+2 X 83 = 2 X

512 =

+ 1024

1143
Therefore, 21 67 8 = 1143,„.

DECIMAL-TO-OCTAL CONVERSION

There are two commonly used methods for converting decimal num-
bers to their octal equivalents. The reader may choose the method
which he prefers.

SUBTRACTION OF POWERS METHOD. The following procedure
is followed to convert a decimal number to its octal equivalent. Sub-
tract from the decimal number the highest possible value of the form
«8 n , where a is a number between 1 and 7, and n is an integer. Record
the value of a. Continue to subtract decreasing powers of 8 (recording
the value of a each time) until the decimal number is reduced to zero.
Record a value of a=0 for all powers of 8 which could not be subtrac-
ted. Table 1-2 may be used to convert any number which can be rep-
resented by 12-bits (4095 10 or less). Appendix F contains a similar
table for converting larger numbers. Example:

2591,„ = ? octal

2591 .

-2560 = 5 X 8 3 = 5 X 512 5 3 7

31 . : T

♦. 4*

- = 0X8 2 = 0~X 64

24 = 3 X 8 ] = 3 X

7 I

- 7 = 7 X 8° = 7 X 1

Therefore, 259 1 10 = 503 7 S .

1-12

ble 1-2. Octs

il-Decimal Conversion

Octal
Digit

Position Coefficients
(Multipliers)

. 3.

1st (8°)

2nd (8 1 )

- 32

3rd (82)

128

192

256-

320

384

448

4th (8 s )

512

1,024

1,536

2,048

2,560

3,072

3,584

DIVISION METHOD. A second method for converting a decimal
number to its octal equivalent is by successive division by 8. Divide the
decimal number by 8 and record the remainder as the least significant
digit of the octal equivalent. Continue dividing by 8, recording the re-
mainders as the successively higher significant digits until the quotient
is reduced to zero. Example:

1376 M = ? octal

Quotient

Remainder

8 ^1376 172

8 ynr 21

8 W 2
8 JT ■

2 5 A

[

Therefore, 1376 M = 2540 8 .

1-13

EXERCISES

a. Convert the following binary numbers to their octal equivalents.

1- 1110 9. 10111111

2- 0110 10. 111111111111

3. HI 11. 010110101011

4. 101111101 12. 111110110100

5. 110111110 13. 010100001011

6. 100000 14. 000010101101

7. 11000111 15. 110100100100
8- 011000 16. 010011111010

b. Convert the following octal numbers to their binary equivalents.

1. 354 9. 70

2. 736 10. 64

3. 15 11. 7777

4. 10 12. 7765

5. 7 13. 3214

6. 5424 14. 4532

7. 307 15. 7033

8. 1101 16. 1243

c. Convert the following decimal numbers to their octal equivalents.

1. 796 7. 1080

2. 32 8. 1344

3. 4037 9. 1512

4. 580 10. 3077

5. 1000 11. 4056

6. 3 12. 4095

d. Convert the following octal numbers to their decimal equivalents.

1. 17 7. 7773

2. 37 8. 7777

3. 734 9. 3257

4. 1000 10. 4577

5. 1200 11. 0012

6. 742 12. 0256

1-14

Fractions

The binary and octal number systems represent fractional parts of
numbers in a similar manner to the decimal system. Furthermore, frac-
tions may be converted from one number system to another by the
same techniques developed for converting whole numbers.

Before investigating the mechanics of fraction conversion, consider
what a fraction is. A fraction is a number between and 1, or a num-
ber less than a unit. Until now only whole numbers in the following
three systems have been considered: decimal, binary, and octal. In
each of these systems, the position of the symbol in the number denotes
its power, and the symbol is the coefficient of that power. These are
positive powers. For example, in the decimal system the number 598,
5 is the coefficient of 10 2 , 9 is the coefficient of 10\ and 8 is the coeffi-
cient of 10°. In binary and octal the same rule applies to using the
powers of the base of the system.

When working with fractions, an important point to keep in mind
is that fractions contain coefficients of negative powers, with the radix
point being the dividing line between the non-negative and negative
powers of the number system being used. Any number to the im-
mediate right of the radix point has a power of negative (minus) 1.
The first digit of the fractional number is the MSD. For example, in
the decimal fraction .637; 6 is the coefficient of 10 1 , 3 is the coeffi-
cient of 10- 2 , and 7 is the coefficient of 10" 3 .- The coefficient of a nega-
tive power of the base, is actually the numerator of a proper fraction
whose denominator is the positive power of that base. For example,
.6 10 (6 x 10 1 ) is equivalent to 6 divided by 10 1 or 6/10, and also
.3 8 (3 x 8" 1 ) is equivalent to 3 divided by 8 1 or 3/8. It should be ap-
parent that this general rule applies to any base that may be considered.
Table 1-3 contains proper fractions which have been changed to' deci-
mal, binary, and octal for comparison purposes.

CONVERTING DECIMAL FRACTIONS TO
BINARY AND OCTAL FRACTIONS

SUBTRACTION OF POWERS METHOD. One method of converting
a decimal fraction to a different number system is the subtraction of
powers method. In this method, subtractions of the highest possible
negative power of a. number in another system that is contained in the
decimal fraction, are performed. In each subtraction, recording the
power and its coefficient gives the equivalent number in the other sys-
tem. When no subtraction is possible, a is recorded. To convert a
decimal fraction to a binary fraction, the powers of 2 are associated

1-15

Table 1-3. Fraction Equivalents

Proper

Decimal

, Octal

Binary

Fraction

Equivalent

1/2

1/4

.25

.20

.01

1/8

.125

.10

.001

1/16

.0625

.04

.0001

1/32

.03125

.02

.00001

1/64

.015625

.01

.000001

1/128

.0078125

.004

.0000001

1/256

.00390625

.002

.00000001

1/512

.001953125

.001

.000000001

1/1024

.0009765625

.0004

.0000000001

with coefficients of or 1, since they are the only coefficients used in
this system. In the octal system, the coefficients through 7 are used.
The following example and explanation will show the conversion of
the decimal fraction .5625 to binary.

.5625
.5000=2-!

.0625
-.0625=2"*

.0625

.0000

Negative Powers of 2
Decimal Equivalents
Bit Values of Answer

2- 1

2-2

2-3

2-*

.5000

.2500

.1250

.0625

.1001

The largest negative power of 2 contained in the decimal fraction .5625
is 2" 1 , which is equivalent to decimal .5000; subtract .5000 10 from
.5625,0 and record a 1 in the 2' 1 column. It is not possible to subtract
2' 2 from the remainder, so record a in the 2' 2 column, 2' 3 cannot be
subtracted from the remainder, so record a in the 2" 3 column; 2~ 4
can be subtracted from the remainder, so record a 1 in the 2"* column.
Thus, the binary equivalent of a decimal .5625 is .1001 2 .

Conversion to octal fractions follows the same procedure, but more
than one subtraction of a given power of the base is possible. The
number of times this subtraction is possible yields the coefficient of that
particular power of the base. This method will not be demonstrated
here, since it is very cumbersome, and easier methods are available.

1-16

MULTIPLICATION METHOD. This method of conversion is fre-
quently used to change from a decimal fraction to another base. To
convert, the decimal fraction is multiplied through by the base of the
system being converted to. For example, convert decimal fraction .5625
to binary. Multiply the decimal fraction by 2. Since a whole number is
obtained, record a 1 in the 2~ x column, discard the whole number por-
tion of the number, and multiply the remainder by 2 again. No whole
number is obtained, so record a in the 2' 2 column, and multiply the
result by 2. No whole number is obtained, so record a in the 2~' A
column, and multiply by 2 again. A whole number is obtained, so record
a 1 in the 2~ 4 column. The remainder, now reduced to 0, completes the
conversion, and .5625 10 is .1001 2 . The following examples show the con-
version just described, and the same decimal fraction converted to octal.

Decimal to Binary Decimal to Octal

.5625 .5625

2 8

1.1250 ( 4.5000

2 8

■ 0.2500 .44 < 4.0000

■0.5000

.1001* 1.0000

CONVERTING BINARY AND OCTAL TO DECIMAL
FRACTIONS

EXPANSION METHOD. This method can be used in converting frac-
tions from any base to a decimal fraction. Remember that the MSD is
the first digit to the right of the radix point in a fractional number, and
that it is multiplied by the base to the —1 power. The second digit is
that digit multiplied by the base to the —2 power, etc. For example, to
convert the binary fraction .10001 to decimal, proceed, as follows. The
MSD is 1 X (2" 1 ) or 1/2, the second digit is X (2~ 2 ) or 0, the third
digit is X (2- 8 ) or 0, the fourth digit is X (2~ 4 ) or 0, and the
fifth digit is 1 X (2~ 5 ) or 1/32. The binary numbers are multiplied by
the respective powers and added together to get the answer. Thus
1/2 + 1/32 which is 16/32 + 1/32 equals 17/32 or .53125 10 .

1-17

The octal fraction A2 can be converted in the same manner, as fol-
lows. The MSD is 4 X (8" 1 ) or 4/8 and 2 X (8~ 2 ) or 2/64. The frac-
tions are now added together to get the result; 4/8 + 2/64 or
16/32 + 1/32 = 17/32 or".53125 10 . If you look carefully at the
binary fraction . 10001 2 and divide it into groups of 3 to convert to
octal, you can see that .10001 2 does equal .42 8 . Zeros may be added
to the right of a fraction without changing the value.
"SHORT CUT" METHOD. This is another method of converting frac-
tions from another base to decimal. In this method, start at the LSD of
the fraction and proceed to the MSD of the fraction, counting the
powers of the base, the next higher power of the base will be utilized
as a common denominator. The number is assumed to be a whole num-
ber for counting purposes. The number .10001 2 would be converted
as follows:

.10 1

2 4 23 2 2 2 1 2°

The MSD is 2 4 or 16, so the common denominator is the next higher
power of 2, or 32. The numerator is converted as if it were a whole
number. The result is then 17/32 which is .53125k,. The same method
with the octal fraction .42 should yield the same result.

.4 2
8 1 8°

The MSD is 8 1 , or 8, so the common denominator is the next higher
power of 8, or 64. Multiplying the digit values by the powers of the
base and adding the products gives us the value of the numerator; thus,
4 X (8 1 ) + 2 X (8°) = 34, and the fraction 34/64 equals .53125 10 .

Arithmetic Operations with Binary and Octal Numbers

Now that the reader understands the conversion techniques between
the familiar decimal number system and the binary and octal number
systems, arithmetic operations with binary and octal numbers will be
described. The reader should remember that the binary numbers are
used in the computer and that the octal numbers are used as a means
of representing the binary numbers conveniently.

BINARY ADDITION

Addition of binary numbers follows the same rules as decimal or
other bases. In adding decimal 1 + 8 we have a sum of 9. This is the
highest value digit. Adding one more requires the least significant digit

1-18

to become a with a carry of 1 to the next place in the number. Simil-
arly, adding binary + 1 we reach "the highest value a single digit can
have in the binary system, and adding one more (1 + 1) requires a
carry to the next higher power (1 + 1 = 10). Take the binary numbers
101 + 10 (5 + 2).

= 7„

+ l = l,l+0 = i, and + 1 = 1 with no carries required. The
answer is 111, which is 7. Suppose we add 111 to 101.

i i < carries

111 = 7 W

+ 101 = 5 M

1100 = 12 10

Now 1 + 1=0 plus a carry of 1 . In the second column, 1 plus the
carry 1 = 0, plus another carry. The third column is 1 + 1 = with
a carry, plus the previous carry, or 1 + 1. + 1 = 1 1 . Our answer 1 100
is eqjiial to 1 X 2 3 + 1 X 2 2 or 8 + 4 = 12, which is the correct
solution for 7 + 5.

OCTAL ADDITION

Addition for octal numbers should be no problem if we keep in mind
the following basic rules for addition.

1. If the sum of any column is equal to or greater than the base
of the system being used, the base must be subtracted from the
sum to obtain the final result of the column.

2. If the sum of any column is equal to or greater than the base,
there will be- a carry to the next column equal to the number
of times the base was subtracted.

3. If the result of any column is less than the base, the base is
not subtracted and no carry will be generated. Examples:

5 8 = 5i<> 3 5 8 = 29 10

+ 3s = 3„ 6 3, = 51 10

~~T 1 10 8

- 8 -8-8

"~T07 =8. 12 8 = 80,„

1-19

Negative Numbers and Subtraction

Up to this point only positive numbers have been considered. Neg-
ative numbers and subtraction can be handled in the binary system in
either of two ways: direct binary subtraction or by the two's comple-
ment method.

BINARY SUBTRACTION (DIRECT)

Binary numbers may be directly subtracted in a manner similar to
decimal subtraction. The essential difference is that if a borrow is re-
quired, it is equal to the base of the system or 2.

110 = 6 10

-101 = 5„

001 = i„

To subtract 1 from in the first column, a borrow of 1 was made from
the second column which effectively added 2 to the first column. After
the borrow, 2 — 1 = 1 in the first column; in the second column
— = 0; and in the third column 1 — 1 = 0. The same numbers
which were subtracted using the twos complement method are sub-
tracted directly in the foljowing example.

Oil 001 100 010 B

010 010 010 111 A

000 111 001 011 B-A

TWO'S COMPLEMENT ARITHMETIC

To see how negative numbers are handled in the computer, consider
a mechanical register, such as a car mileage indicator, being rotated
backwards. A 5-digit register approaching and passing through zero
would read the following.

00005
00004
00003
00002
00001
00000
99999
99998
etc.

1-20

It should be clear that the number 99998 corresponds to —2. Fur-
ther, if we add

00005
99998

1 00003

and ignore the carry to the left, we have effectively performed the

operation of subtracting

5-2=3

The number 99998 in this example is described as the ten's complement
of 2. Thus in the decimal number system, subtraction may be per-
formed by adding the ten's complement of the number to be subtracted.

If a system of complements were to be used for representing negative
numbers,- the minus sign could be omitted in negative numbers. Thus
all numbers could be represented with five digits; 2 represented as
00002, and —2 represented as 99998. Using such a system requires
that a convention be established as to what is and is not a negative
number. For example, if the mileage indicator is turned back to 48732,
is it a negative 51268, or a positive 48732? With an ability to represent
a total of 100,000 different numbers (0 to 99999),. it would seem
reasonable to use half for positive numbers and half for negative num-
bers. Thus, in this situation, to 49999 would be regarded as positive,
and 50000 to 99999 would be regarded as negative.

In this same manner, the two's complement of binary numbers are
used to represent negative numbers, and to carry out binary subtraction,
in the PDP-8 computer. In octal notation, numbers from 0000 to 3777
are regarded as positive and the numbers from 4000 to 7777 are re-
garded as negative.

The two's complement of a number is defined as that number which
when added to the original number will result in a sum of unity. The
binary number 110110110110 has a two's complement equal to
001001001010 as shown in the following addition.

110 110 110 110
001 001 001 010

1 000 000 000 000

The easiest method of finding a two's complement is to first obtain the
one's complement, which is formed by setting each bit to the opposite
value.

101 000 110 111 Number

010 111 001 000 One's complement of the number

1-21

The two's complement of the number is then obtained by adding 1 to
the one's complement.

110 001 110 010 Number

001 110 001 101 One's complement of the number

+1_ Addl

001 110 001 110 Two's complement of the number

Subtraction in the PDP-8 is performed using the two's complement
method. That is, to subtract A from B, A must be expressed as its two's
complement and then the value of B is added to it. Example:

010 010 010 111 A

101 101 101 001 Two's complement of A

(carry is 011 001 100 010 B

ignored) 1 000 111 001 011 B-A

OCTAL SUBTRACTION

Subtraction is performed in the octal number system in two ways
which are directly related to the subtractions in the binary system. Sub-
traction may be performed directly or by the radix (base) complement
method.

OCTAL SUBTRACTION (DIRECT). Octal subtraction can be per-
formed directly as illustrated in the following examples.

3567-2533= ? 2022-1234== ?

3567 2022

-2533 -1234

1034 0566

Whenever a borrow is needed in octal subtraction, an 8 is borrowed as
in the second example above. In the first column, an 8 is borrowed
which is added to the 2 already in the first column and the 4 is sub-
tracted from the resulting 1 0. In the second column, an 8 is borrowed
and added to the 1 which is already in the column (after the previous
borrow) and the 3 is subtracted from the resulting 9. In the third
column the 2 is subtracted from a borrowed 1 (originally a borrowed
8 ) , and in the last column 1—1 =0.

EIGHT'S COMPLEMENT ARITHMETIC. Octal subtraction may be
performed by adding the eight's complement of the subtrahend to the
minuend. The eight's complement is obtained in the following manner.

3042 Number <

4735 Seven's complement of the number
+1 Add 1 to seven's complement to obtain

4736 Eight's complement

1-22

The seven's complement of the number is obtained by setting each digit
of the complement to the value of 7 minus the digit of the number, as
seen above. The eight's complement of the number is then obtained by
adding 1 to the seven's complement. To prove that the complement is
in fact a complement, the number is added to the complement and a re-
sult of zero and an overflow of 1 is obtained.

3042

+4736

1 0000

The following example uses the eight's complement to subtract a
number.

3567-2533=

2533

Number

5244

Seven's complement

5245

Eight's complement

(carry is

3567

+5245

Minuend

Eight's complement of subtrahend

ignored) ?

►1 1034

Difference

Multiplication and Division in Binary and Octal Numbers

Though multiplication in computers is usually achieved by means
other than formal multiplication, a formal method will be demonstrated
as a teaching vehicle.

BINARY MULTIPLICATION

In binary multiplication, the partial product is moved one position to
the left as each successive multiplier is used. This is done in the same
manner as in decimal multiplication. If the multiplier is a 0, the partial
product can be a series of 0s as in example 2, or the next partial product
can be moved two places to the left as in example 3, or three places as
in example 4.

Example 1.

462 1(
127 M

3234
924
462

58674

Multiplicand
Multiplier

First partial product
Second partial product
Third partial product

Product

1-23

Example 2.

Example 3.

1110110,

ion,

mono
mono
ooooooo
mono

10100010010,

mono,

101L

Example 4.

mono
mono
mono

10100010010,

11001110,
11001,

11001110
11001110

nooino

1010000011110,

Because of the difficult binary additions resulting from multiplica-
tions such as the previous examples, octal multiplication of the octal
equivalents of binary numbers is often substituted.

OCTAL MULTIPLICATION

Multiplication of octal numbers is the same as multiplication of
decimal numbers as long as the result is less than 10 8 . Obviously this
could be a problem if it weren't for the fact that an octal multiplication
table can be set up, similar to the decimal multiplication table, to make
the job of multiplication of octal numbers quite simple. Table 1-4
is a partially completed octal multiplication table that will be quite use-
ful once you have filled in the blank squares.

Using the completed octal multiplication table, the following prob-
lems may be solved.

226 s X 12, = ?

2734.

1-24

1247„*305 8 =?

1247 8
X305 8

6503
0000
3765

405203 8

Table 1-4. Octal Multiplication Table

BINARY DIVISION

Once the reader has mastered binary subtraction and multiplication,
binary division is easily learned. The following problem solutions illus-
trate binary division.

Divide 10010 2

10 a
1001

10 } 10010
2£

00
00

01
00

10
10

10010 2

ioT

18 10

= 1001 a

1-25

Divide 1110 a

14*

100 2 ~

^10

11.1

100 5 ino.o

100

110
100

100
100

3.5„

ILL = 3.5!,

OCTAL DIVISION

Octal division uses the same principles as decimal division. All mul-
tiplication and subtraction must however be done in octal. (Refer to the
octal multiplication table.) The following problem solutions illustrate
octal division.

62„

2 8

2 562"

1714.

22 s

= 31.

= 25 1(

EXERCISES

a. Perform the following binary additions.

10110

101

10.

100111

+101

1
+ 110

111001
+ 101101

100

•

+10

1110
100

11.

11011001
10010011

11011

+11

+11100011

+0010

12.

11011011

10110111

101

10111011

+ 1

+ 1000

00101011
01010111

1101

110111

+01111101

101

100100

+11 ■

+ 110001

1-26

b. Find the one's complement and the two's complement of the fol-

wing numoers.
Oil 100 110 010

000 000 000 111

010 111 011 111

100 000 000 000

Oil 110 000 000

100 000 010 010

000 000 000 000

10.

100 001 100 110

000 000 000 001

11.

111 111 111 110

000 100 100 100

12.

111 111 111 111

c. Subtract the following binary numbers directly.

1. 101000001
010111101

1010111010
0101110101

101011010111
011111111101

101111100111
010101110010

d. Perform the following subtractions by the two's complement
method. Check your work by direct subtraction. Show all work.

1. Oil Oil Oil Oil — 001 111 010 110

2. 000 111 111 111 — 000 001 001 101

3. Oil 111 111 101 — 010 101 100 Oil

4. 001 101 111 110 — 001 100 101 Oil

5. Oil 111 111 111 — 010 101 101 101

e. Multiply the following binary numbers.

1. 11011
X110

1011101
X101

101011101011
X 10000

f . Divide the following binary numbers.

100
10

10000

1100100

100

10100

1-27

127

777

256

543

+724

+612

437

+ 11

426

772

3357

747

+562

+575

2543

-44

-2174

7474

7500

—4777

-6373

7000

-6573

g. Add the following octal numbers.

1. 42 6.

+53

2. 45

+23

3. 34 5.

+76

h. Subtract the following octal numbers directly.

1. 42 4.
-23

2. 76 5.

-34

3. 77 6.
-11

i. Perform the following octal subtractions by the eight's comple-
ment method. Check your work by subtracting directly. Show all
work.

1. 0377 - 0233 5. 2311 - 2277

2. 2345 - 1456 6. 0044 - 0017

3. 1144 - 1046 7. 3234 - 2777

4. 3000 - 0011 8. 1111 — 0777

j. Multiply the following octal numbers.

1. 65 3. 77 5. 425
X4 X65 X377

2. 14 4. 716 6. 571
Xl3 X472 X246

k. Prove the answers to the problems in (j) by division, as follows:

Multiplicand

X Multiplier Multiplicand

Product Multiplier!) Product

1-28

LOGIC OPERATION PRIMER

Computers use logic operations in addition to arithmetic operations
to solve problems. The logic operations have a direct relationship with
the algebraic system to represent logic statements known as Boolean
algebra. In logic, there are two basic connectives that are used to ex-
press the relationship between two statements. These are the AND and
the OR.

The AND Operation

The following simple circuit with two switches illustrates the AND
operation. If current is allowed to flow through a switch, the switch is
said to have a value of 1 . If the switch is open and current cannot flow,
the switch has a value of 0. If the whole circuit is considered, it will
have a value of 1 (i.e., current may flow through it) whenever both A
and B are 1 . This is the AND operation.

^rt

The AND operation is often stated A • B = F. The multiplication sym-
bol (•) is used to represent the AND connective. The relationship be-
tween the variables and the resulting value of F is summarized in the
following table.

When the AND operation is applied to binary numbers, a binary 1 will
appear in the result if a binary 1 appeared in the corresponding position
of the two numbers.

The AND operation can be used to mask out a portion of a 12-bit
number.

To Be To Be Retained

Masked for Subsequent
Out Operation

010 101 010 101
000 000 111 111
000 000 010 101

(12-bit number)

(mask)

(result)

1-29

The OR Operation

A second logic operation is the OR (sometimes called the inclusive
OR). Statements which are combined using the OR connective are
illustrated by the following circuit diagram.

-<r o-

Current in the above diagram may flow whenever either A or B (or
both) is closed (F=l if A=l,orB=l, or A=l and B= 1 ). This opera-
tion is expressed by the plus (+) sign; thus A+B=F. The following
table shows the resulting value of F for changing values of A and B.

Thus, if A and B are the 12-bit numbers shown below, A+B is eval-
uated as follows.

A = Oil 010 Oil 111
B = 100 110 010 011

A + B = 111 110 011 111

Remember that the "+" in the above example means "inclusive OR",
not "add."

The Exclusive OR Operation

The third and last logic operation is the exclusive OR. The exclusive
OR is similar to the inclusive OR with the exception that one set of
conditions for A and B are excluded. This exclusion can be symbolized
in the circuit diagram by connecting the two switches mechanically to-
gether. This connection makes it impossible for the switches to be closed

1-30

simultaneously, although they may be open simultaneously or individu-
ally.

Thus, the circuit is completed when A=l and B=0, and when A=0
and B=l. The results of the exclusive OR operation are summarized
in the table below.

The exclusive OR of two 12-bit numbers is evaluated and labeled F
in the following operation.

A= Oil 010 Oil 111
B = 100 110 010 011

F = 111 100 001 100

GENERAL ORGANIZATION OF THE PDP-8

Almost every general purpose digital computer has the basic units
shown in Figure 1-1, on the following page.

If a machine is to be called a computer, it must have the capability of
performing some types of arithmetic operations. The element of a digital
computer that meets this requirement is called the arithmetic unit. In
order for the arithmetic unit to be able to do its required task, it must
be told what to do. Therefore, a control unit is necessary.

Since mathematical operations are performed by the arithmetic unit,
it may be necessary to store a partial answer while the unit is computing
another part of the problem. This stored partial answer can then be used
to solve other parts of the problem. It is also helpful for the control unit
and arithmetic unit to have information immediately available for their
use, and for the use of other units within the computer. This require-
ment is met by the portion of the computer designated as the memory
unit, or core storage urtit.

1-31

1 1 —
1

1
1

CONTROL
UNIT

1
1

INPUT
UNIT

INTERNAL
STORAGE

OUTPUT
UNIT

ARITHMETIC
UNIT

Figure 1-1. PDP-8 General Organization

The prime purpose of a digital computer is to serve humans in some
manner. In order to do this there must be a method of transmitting our
wants to the computer, and a means of receiving the results of the com-
puter's calculations. The portions of the computer that carry out these
functions are the input and output units.

Arithmetic Unit

The arithmetic unit of a digital computer performs the actual work
of computation and calculation. It carries out its job by counting series
of pulses or by the use of logic circuits. Modern computers use com-
ponents such as transistors and integrated circuits. Switches and relays
were used previously, and were acceptable as far as their ability to per-
form computations was concerned. Modern computers, however, be-
cause of the speed desired, make use of smaller electronic components
whenever possible.

The arithmetic unit of the PDP-8 has, as its major component, a
12-bit accumulator, which is simply a register capable of storing a num-
ber of 12 binary digits. It is called the accumulator because it accumu-
lates partial sums during the operation of the PDP-8. All arithmetic
operations are performed in the accumulator of the PDP-8.

Control Unit

The control unit of a digital computer is an administrative or switch-
ing section. It receives information entering the machine and decides
how and when to perform operations. It tells the arithmetic unit what
to do and where to get the necessary information. It knows when the
arithmetic unit has completed a calculation and it tells the arithmetic
unit what to do with the results, and what to do next.

1-32

The control unit itself knows what to tell the arithmetic unit to do by
interpreting a set of instructions. This set of instructions for the control
unit is called a program and is stored in the computer memory.

Memory Unit

The memory unit, sometimes called the core storage unit, contains
information for the control unit (instructions) and for the arithmetic
unit (data). The terms core storage and memory may be used inter-
changeably. Some computer texts refer to external units as storage, such
as magnetic tapes and disks, and to internal units as memory, such as
magnetic cores. The requirements of the internal storage units may vary
greatly from computer to computer.

The PDP-8 memory unit is composed of magnetic cores which are
often compared to tiny doughnuts. These magnetic cores record binary
information by the direction in which they are magnetized (clockwise
or counterclockwise). The memory unit is arranged in such a way that
it can store 4096 "words" of binary information. These words are each
12-bits in length. Each core storage location has an address, which is a
unique number used by the control unit to specify that location. Storage
of this type in which each location can be specified and reached as easily
as any other is referred to as random-access storage. The other type of
storage is sequential storage such as magnetic tape, in which case some
locations (those at the beginning of the tape) are easier to reach than
others (those at the end of the tape) .

Input Unit

Input devices are used to supply the values needed by the computer
and the instructions to tell the computer how to operate on the values.
Input unit requirements vary greatly from machine to machine. A
manually operated keyboard may be sufficient for a small computer.
Other computers requiring faster input use punched cards for data in-
puts. Some systems utilize removable plugboards that can be pre-wired
to perform certain instructions. Input may also be via punched paper
tape or magnetic tape, two forms of input common in PDP-8 systems.

Output Unit

Output devices record the results of the computer operations. These
results may be recorded in a permanent form (e.g., as a printout on the
teleprinter) or they may be used to initiate a physical action (e.g., to
adjust a pressure valve setting) . Many of the media used for input, such
as paper tape, punched cards, and magnetic tape, can also be used for
output.

1-33

COMPUTER DATA FORMATS

The PDP-8 uses 1 2-bit words to represent data. Some of the formats
in which this data is represented are described in the following para-
graphs.

Alphabetic Characters

Computers are designed to operate upon the binary numbers which
it conveniently represents with electronic components. There are occa-
sions however when it is desirable to have the computer represent
characters of the alphabet and punctuation marks. Binary codes are
used to represent such characters. For example, the reader is familiar
with punched cards, which use a system of punched holes to represent
information. Each of these codes associates some character with a par-
ticular binary number. The computer can store the binary number (not
the character) in its memory. When so directed, the computer will out-
put the binary code to a device which will interpret the code and print
the character. Some specific binary codes used to represent alpha-
numeric information (letters, numbers, and punctuation symbols) are
presented in Appendix B.

Number Representations

The PDP-8 operates upon 12-bit words (namely Oto 111 111 111
Ills, or to 7777 8 ). By convention, one half of the numbers are con-
sidered positive (0 to Oil 111 111 111 2 , or to 3777 s ), and one
half (100 000 000 000 2 to 111 111 111 111 2 or 4000 8 to 7777 8 ) are
considered negative. Therefore the PDP-8 can directly represent the
portion of the number line shown in Figure 1-2.

4000„

■"""V r 7^ r

40OI„ 6000 8 / \ 2000 e

<-3777„) (-2000 8 )

7777
(-D

Figure 1-2. PDP-8 Octal Number Line

Notice that the first digit of the 12-bit binary numbers is in effect a
"sign bit." That is, bit (the first bit) specifies the sign of the number
by the following rule. If bit is a 0, the number is positive; if bit is
a 1, the number is negative. This is the means by which the computer

1-34

tests for positive and negative numbers. Thus, the zero is considered

positive. In figure 1-2 it should be noted that the number 4000 is

peculiar in that it has no positive counterpart. (Expressed in octal, the

"two's complement" of 3776 is 4002; of 3777 is 4001; of 4000 is

4000.)

When the octal to decimal conversions are performed, the number

line of Figure 1-2 is converted to the number line of Figure 1-3. Thus

the PDP-8 can represent directly the numbers between — 2048i„ and

+2047 10 . This would seem to be a serious restriction. Through two

techniques however this limitation is overcome.

f f 1

-2048 +2047

Figure 1-3. PDP-8 Decimal Number Line

DOUBLE PRECISION NUMBERS

The PDP-8 memory is made up of 12-bit storage locations. Suppose
however that a number larger than 12-bits were to be stored. By using
two 12-bit storage locations, numbers between — 8,388,608 1( , and
8,388,607 10 may be represented directly. This method of representation
is appropriately called double precision. The method could be extended
to triple precision and further if necessary.

It should be noted that to add double precision numbers, two addi-
tions are needed. Double precision arithmetic is described in Chapter 3.

FLOATING POINT NUMBERS

Another m6thod of representing numbers in the PDP-8 with more
than one 12-bit word is floating point notation. In this notation, a
number is divided into two parts, namely a mantissa (number part)
and an exponent (to some base). In the decimal number system for
example, the number 12 can be written in the following ways.

MANTISSA '

EXPONENT

.12

10 2

1.2

10 1

12.

10°

120.

io- 1

1200.

10-2

PDP-8 floating point notation makes use of a representation similar to the
above with the exception that the exponent and the mantissa are binary

1-35

numbers. The binary mantissa (number part) is stored in two locations
and a third location stores the exponent. The exponent is selected such
that the mantissa has no leading zeros, thereby retaining the maximum
number of significant digits. Further description of the floating point
system is contained in Chapter 6.

1-36

Chapter 2

Programming
Fundamentals

This chapter describes the three general types of computer instruc-
tions and the way in which they are used in computer programs. The
first type of instruction is distinguished by the fact that it operates upon
data that is stored in some memory location and must tell the com-
puter where the data is located in core so that the computer can
find it. This type of instruction is said to reference a location in core
memory; therefore, these instructions are often called memory reference
instructions (MRI).

When speaking of memory locations, it is very important that a clear
distinction is made between the address of a location and the contents
of that location. A memory reference instruction refers to a location by
a 12-bit address; however, the instruction causes the computer to take
some specified action with the content of the location. Thus, although
the address of a specific location in memory remains the same, the con-
tent of the location is subject to change. In summary, a memory refer-
ence instruction uses a 12-bit address value to refer to a memory
location, and it operates on the 12-bit binary number stored in the
referenced memory location.

The second type of instructions are the operate microinstructions,
which perform a variety of program operations without any need for
reference to a memory location. Instructions of this type are used to
perform the following operations: clear the accumulator, test for neg-
ative accumulator, halt program execution, etc. Many of these operate
microinstructions can be combined (microprogrammed) to increase the
operating efficiency of the computer.

The third general type of instructions are the input/output transfer
(IOT) instructions. These instructions perform the transfer of infor-
mation between a peripheral device and the computer memory. IOT
instructions are discussed in Chapter 5.

2-1

PROGRAM CODING

Binary numbers are the only language which the computer is able
to understand. It stores numbers in binary and does all its arithmetic
operations in binary. What is more important to the programmer, how-
ever, is that in order for the computer to understand an instruction it
must be represented in binary. The computer can not understand in-
structions which use English language words. All instructions must be in
the form of binary numbers (binary code) .

Binary Coding

The computer has a set of instructions in binary code which it "un-
derstands". In other words, the circuitry of the machine is wired to react
to these binary numbers in a certain manner. These instructions have
the same appearance as any other binary number; the computer can
interpret the same binary configuration of O's and l's as data or as an
instruction. The programmer tells the computer whether to interpret
the binary configuration as an instruction or as data by the way in which
the configuration is encountered in the program.

Suppose the computer has the following binary instruction set.

Instruction A 001 000 010 010 This binary number instructs the

computer to add the contents of
location 000 000 010 010 to the
accumulator.

Instruction B 001 000 010 111 This binary number instructs the

computer to add the contents of
location 000 000 010 111 to the
accumulator.

If instruction B is contained in a core memory location with an
address of 000 000 010 010 and the binary number 000 111 111 111
is stored in a location with an address of 000 000 010 1 1 1, the follow-
ing program could be written:

Location Content

000 000 010 010 001 000 010 111

000 000 010 111 000 111 111 111

If this program were to be executed, the number 000 111 111 111
would be added to the accumulator.

Octal Coding

If binary configurations appear cumbersome and confusing, the
reader will now understand why most programmers seldom use the
binary number system in actual practice. Instead, they substitute the

2-2

octal number system which was discussed in Chapter 1. The reader
should not proceed until he understands these two number systems
and the conversions between them.

Henceforth, octal numbers will be used to represent the binary num-
bers which the computer uses. Although the programmer may use octal
numbers to describe the binary numbers within the computer, it should
be remembered that the octal representation itself does not exist within
the computer.

When the conversion to octal is performed, Instruction B becomes
1027 8 and the previous program becomes.

Location Content

0022 s 1027 8

0027„ 0777»

To demonstrate that a computer cannot distinguish between a num-
ber and an instruction, consider the following program.

Location Content

0021 - 1022 (Instruction A)

0022 1027 (Instruction B)

0027 0777 (The number 777 s )

Instruction A, which adds the contents of location 0022 to the accu-
mulator, has been combined with the previous program. Upon execu-
tion of the program (assuming the initial accumulator value=0), the
computer will execute instruction A and add 1027 8 as a number to the
accumulator obtaining a result of 1027 s . The computer will then execute
the next instruction, which is 1027, causing the computer to add the
contents of 0027 to the accumulator. After the execution of the two
instructions the number 2026 s is in the accumulator. Thus, the above
program caused the number 1027 s to be used as an instruction and as
a number by the computer.

Mnemonic Coding

Coding a program in octal numbers, although an improvement upon
binary coding, is nevertheless very inconvenient. The programmer must
learn a complete set of octal numbers which have no logical con-
nection with the operations they represent. The coding is difficult- for
the programmer when he is writing the program, and this difficulty is
compounded when he is trying to debug or correct a program. There is
no easy way to remember the correspondence between an octal number
and a computer operation.

2-3

To simplify the process of writing or reading a program, each in-
struction is often represented by a simple 3- or 4-letter mnemonic
symbol. These mnemonic symbols are considerably easier to relate to a
computer operation because the letters often suggest the definition of
the instruction. The programmer is now able to write a program in a
language of letters and numbers which suggests the meaning of each
instruction.

The computer still does not understand any language except binary
numbers. Now, however, a program can be written in a symbolic lan-
guage and translated into the binary code of the computer because of
the one-to-one correspondence between the binary instructions and the
mnemonics. This translation could be done by hand, defeating the pur-
pose of mnemonic instructions, or the computer could be used to do the
translating for the programmer. Using a binary code to represent alpha-
betic characters as described in Chapter 1, the programmer is able to
store alphabetic information in the computer memory. By Instructing
the computer to perform a translation, substituting binary numbers for
the alphabetic characters, a program is generated in the binary code
of the computer. This process of translation is called "assembling" a
program. The program that performs the translation is called an
assembler.

Although the assembler is described in detail in Chapter 6, it is well
to make some observations about the assembler at this point.

1. The assembler itself must be written in binary code, not
mnemonics.

2. It performs a one-to-one translation of mnemonic codes into
binary numbers.

3. It allows programs to be written in a symbolic language which
is easier for the programmer to understand and remember.

A specific mnemonic language for the PDP-8, called PAL (Program
Assembly Language), is introduced later in this chapter. The next sec-
tion describes the general PDP-8 characteristics and components. This
information is necessary to an understanding of the PDP-8 instructions
and their uses within a program.

PDP-8 ORGANIZATION AND STRUCTURE

The PDP-8 is a high-speed, general purpose digital computer which
operates on 12-bit binary numbers. It is a single-address parallel
machine using two's complement arithmetic. It is composed of the five
basic computer units which were discussed in Chapter 1. The com-

2-4

ponents of the five units and their interrelationships are shown in
Figure 2-1. For simplicity, the input and output units have been
combined.

INPUT/

OUTPUT .
UNITS

CONSOLE

INPUT/
OUTPUT
DEVICES

TELE-
TYPE,
DISK,
DEC TAPE,

ETC.

ARITHMETIC
UNIT

LINK

ACCUM-
ULATOR

•4-

CONTROL

UNIT

MEMORY
UNIT

PROGRAM

COUNTER

INSTRUCTION
REGISTER

MAJOR

STATE
GENERATOR

MEMORY

BUFFER
REGISTER

MEMORY
ADDRESS

CORE MEMORY
4096 12-BIT WORDS

Figure 2- 1 . Block Diagram of the PDP-8

Input and Output Units

The input and output units are combined in Figure 2-1 because in'
many cases the same device acts as both an input and an output unit.
The Teletype console, for example, can be used to input information
which will be accepted by the computer, or it can accept processed in-
formation and print it as output. Thus, the two units of input and output
are very often joined and referred to as input/output or simply I/O.
Chapter 5 describes the methods of transmitting data as either input or
output; but for the present, the reader can assume that the computer is
able to accept information from devices such as those listed in the block
diagram and to return output information to the devices. The PDP-8
console allows the programmer direct access to core memory and the
program counter by setting a series of switches, as described in detail
in Chapter 4.

Arithmetic Unit

The second unit contained in the PDP-8 block diagram is the arith-
metic unit. This unit, as shown in the diagram, accepts data from input
devices and transmits processed data to the output devices as well. Pri-
marily, however, the unit performs calculations under the direction of
the control unit. The Arithmetic Unit in the PDP-8 consists of an
accumulator and a link bit.

2-5

ACCUMULATOR (AC)

The prime component of the arithmetic unit is a 12-bit register called
the accumulator. It is surrounded by the electronic circuits which per-
form the binary operations under the direction of the control unit. Its
name comes from the fact that it accumulates partial sums during the
execution of a program. Because the accumulator is only twelve bits in
length, whenever a binary addition causes a carry out of the most sig-
nificant bit, the carry is lost from the accumulator. This carry is re-
corded by the link bit.

LINK (L)

Attached logically to the accumulator is a 1-bit register, called the
link, which is complemented by any carry out of the accumulator. In
other words, if a carry results from an addition of the most significant
bit in the accumulator, this carry results in a link value change from
to 1 , or 1 to 0, depending upon the original state of the link.

Below is a diagram of the accumulator and link. The twelve bits of
the accumulator are numbered to 11, with bit being the most sig-
nificant bit. The bits of the AC and L can be either binary O's or l's as
shown below.

LINK

ACCUMULATOR
4 5 6 7

0/
/I

0/
/ 1

o/ 1

°/°A

MOST SIGNIFICANT BIT

LEAST SIGNIFICANT BIT

Control Unit

The instruction register, major state generator, and program counter
can be identified as part of the control unit. These registers keep track
of what the computer is now doing and what it will do next, thus
specifying the flow of the program from beginning to end.
PROGRAM COUNTER (PC)

The program counter is used by the PDP-8 control unit to record
the locations in memory (addresses) of the instructions to be executed.
The PC always contains the address of the next instruction to be exe-
cuted. Ordinarily, instructions are stored in numerically consecutive
locations and the program counter is set to the address of the next in-
struction to be executed merely by increasing itself by 1 with each
successive instruction. When an instruction causing transfer of command
I to another portion of the stored program is encountered, the PC is set

2-6

to the appropriate address. The PC must be initially set by input to
specify the starting address of a program, but further actions are con-
trolled by program instructions.
INSTRUCTION REGISTER (IR)

The 3-bit instruction register is used by the control unit to specify
the main characteristics of the instruction being executed. The three
most significant bits of the current instruction are loaded into the IR
each time an instruction is loaded into the memory buffer register from
core memory. These three bits contain the operation code which
specifies the main characteristics of an instruction. The other details
are specified by the remaining nine bits (called the operand) of the
instruction.
MAJOR STATE GENERATOR *

The major state generator establishes the proper states in sequence
for the instruction being executed. One or more of the following three
major states are entered serially to execute each programmed instruc-
tion. During a Fetch state, an instruction is loaded from core memory,
at the address specified by the program counter, into the memory
buffer register. The Defer state is used in conjunction with indirect ad-
dressing to obtain the effective address, as discussed under "Indirect
Addressing" later in this chapter. During the Execute state, the instruc-
tion in the memory buffer register is performed.

Memory Unit

The PDP-8 basic memory unit consists of 4,096 12-bit words of
magnetic core memory, a 12-bit memory address register, and a 12-bit
memory buffer register. The memory unit may be expanded in units of
4,096 words up to a maximum of 32,768 words;

CORE MEMORY

The core memory provides storage for the instructions to be per-
formed and information to be processed. It is a form of random access
storage, meaning that any specific location can be reached in memory
as readily as any other. The basic PDP-8 memory contains 4,096 12-bit
magnetic core words. These 4,096 words require that 12-bit addresses
be used to specify the address for each location uniquely.
MEMORY BUFFER REGISTER (MB)

All transfers of instructions or information between core memory and
the processor registers (AC, PC, and IR) are temporarily held in the
memory buffer register. Thus, the MB holds all words that go into and
out of memory, updates the program counter, sets the instruction
register, sets the memory address register, and accepts information
from or provides information to the accumulator.

2-7

MEMORY ADDRESS REGISTER (MA)

The address specified by a memory reference instruction is held in
the memory address register. It is also used to specify the address of the
next instruction to be brought out of memory and performed. It can be
used to directly address all of core memory. The MA can be set by
the memory buffer register, or by input through the program counter
register, or by the program counter itself.

MEMORY REFERENCE INSTRUCTIONS

The standard set of instructions for the PDP-8 includes eight basic
instructions. The first six of these instructions are introduced in the
following paragraphs and are presented in both octal and mnemonic
form with a description df the action of each instruction.

The memory reference instructions (MRI) require an operand to
specify the address of the location to which the instruction refers. The
manner in which locations are specified for the PDP-8 is discussed in
detail under "Page Addressing" later in this chapter. In the following
discussion, the first three bits (the first octal digit) of an MRI are used
to specify the instruction to be performed. (The last nine bits, three
octal digits, of the 12-bit word are used to specify the address of the
referenced location — that is, the operand.)

The six memory reference instructions are listed below with their
mnemonic and octal equivalents as well as their memory cycle times.

Instruction

Logical AND

Two's Complement Add

Deposit and Clear the Accumulator

Jump

Increment and Skip if Zero

Jump to Subroutine

Octal

Memory

Mnemonic 3

Value

Cycles 1

AND

Onnn

TAD

lnnn

DCA

3nnn

JMP

5nnn

ISZ

2nnn

JMS

4nnn

1 Memory cycle time for the PDP-8 and -8/1 is 1.5 microseconds; for the PDP-
8/L, it is 1.6; for the PDP-8/S, it is 8 microseconds. (Indirect addressing re-
quires an additional memory cycle.)

2 The mnemonic code is meaningful to and translated by an assembler into

binary code.

2-8

AND (Onmi 8 )

The AND instruction causes a bit-by-bit Boolean AND operation
between the contents of the accumulator and the data word specified
by the instruction. The result is left in the accumulator as illustrated
below.

LINK

INK 10

DATA WORD

I I AC (RESULT)

The following points should be noted with respect to the AND

instruction: .

1 A 1 appears in the AC only when a 1 is present in both the AC
and the data word (The data word is often referred to as a

in z\ *i lc I "

2. The state of the link bit is not affected by the AND instruction;

and

3. The data word in the referenced location is not altered.

TAD (lnnn 8 )

The TAD instruction performs a binary addition between the speci-
fied data word and the contents of the accumulator, leaving the result
of the addition in the accumulator. If a carry out of the most significant
bit of the accumulator should occur, the state of the link bit is comple-
mented. The add instruction is called a Two's Complement Add to re-
mind the programmer that negative numbers must be expressed as the
two's complement of the positive value. The following figure illustrates
the operation of the TAD instruction.

INK I 10 1 AC:+5

DATA WORD! -3

LINK

AC (RESULT): +2

2-9

The following points should be remembered when using the TAD
instruction:

1. Negative numbers must be expressed as a two's complement of
the positive value of the number;

2. A carry out of the accumulator will complement the link; and

3. The data word in the referenced location is not affected.
DCA (3nnn s )

The DCA instruction stores the contents of the AC in the referenced
location, destroying the original contents of the location. The AC is
then set to all zeroes. The following example shows the contents of the
accumulator, link, and location 225 before and after executing the in-
struction DCA 225.
DCA 225

AC Link Loc. 225

Before Execution 1234 1 7654

After Execution 0000 1 1234

The following facts should be kept in mind when using the DCA in-
struction:

1. The state of the link bit is not altered;

2. The AC is cleared; and

3. The original contents of the addressed location are replaced by
the value of the AC.

JMP (5min 8 )

The JMP instmction loads the effective address of the instruction
into the program counter, thereby changing the program sequence since
the PC specifies the next. instruction to be performed. In the following
example, execution of the instruction in location 250 (JMP 300) causes
the program to jump over the instructions in locations 251 through 277
and immediately transfer control to the instruction in location 300.
Location Content

250 JMP 300 (This instruction transfers program

control to location 300.)

300 DCA 330

NOTE: The JMP instruction does not affect the contents of
the AC or link.

ISZ (2nnn 8 )

The ISZ instruction adds a 1 to the referenced data word and then
examines the result of the addition. If a zero result occurs, the instruc-
tion following the ISZ is skipped. If the result is not zero, the instruction

2-10

following the ISZ is performed. In either case, the result of the addition
replaces the original data word in memory. The example in Figure 2-2
illustrates one method of adding the contents of a given location to the
AC a specified number of times (multiplying) by using an ISZ instruc-
tion to increment a tally. The effect of this example is to multiply the
contents of location 275 by 2. (To add the contents of a given location
to the AC twice, using the ISZ loop, as shown in Figure 2-2, requires
more instructions than merely repeating the TAD instruction. However,
when adding the contents four or more times, use of the ISZ loop re-
quires fewer instructions.) In the first pass of the example, execution of
ISZ 250 increments the contents of location 250 from 7776 to 7777
and then transfers control to the following instruction (JMP 200). In
the second pass, execution of ISZ 250 increments the contents of loca-
tion 250 from 7777 to 0000 and transfers control to the instruction in
location 203, skipping over location 202.
CODING FOR ISZ LOOP
Location Content

200

201
202
203

TAD 275
ISZ 250
JMP 200
DCA 276

250

7776

275 0100

276 0000

SEQUENCE OF EXECUTION FOR ISZ LOOP

Content After Instruction Execution

Location Content

250

275

276

■i ' ■

i> '

FIRST PASS

200 TAD 275

0100

7776

0100

0000

201 ISZ 250

0100

7777

0100

0000

202 JMP 200

0100

7777

0100

0000

SECOND PASS

200 TAD 275

0200

7777

0100

0000

201 ISZ 250

0200

0000

0100

0000

202 JMP 200

(Skipped

during second pass)

203 DCA 276

0000

0100

0200

Figure 2-2. ISZ Instruction Incrementing a Tally

2-11

The following points should be kept in mind when using the 1SZ
instruction :

. 1 . The contents of the AC and link are not disturbed;

2. The original word is replaced in main memory by the incre-
mented value;

3. When using the ISZ for looping a specified number of times,
the tally must be set to the negative of the desired number; and

4. The ISZ performs the incrementation first and then checks for
a zero result.

JMS (4nnn s )

A program written to perform a specific operation often includes sets
of instructions which perform intermediate tasks. These intermediate
tasks may be finding a square root, or typing a character on a keyboard.
Such operations are often performed many times in the running of one
program and may be coded as subroutines. To eliminate the need of
writing the complete set of instructions each time the operation must be
performed, the JMS (jump to subroutine) instruction is used. The JMS
instruction stores a pointer address in the first location of the subroutine"
and transfers control to the second location of the subroutine. After the
subroutine is executed, the pointer address identifies the next instruc-
tion to be executed. Thus, the programmer has at his disposal a simple
means of exiting from the normal flow of his program to perform an
intermediate task and a means of return to the correct location upon
completion of the task. (This return is accomplished using indirect ad-
dressing, which is discussed later in this chapter.) The following exam-
ple illustrates the action of the JMS instruction.

Location Content

PROGRAM

200 JMS 350 (This instruction stores 0201 in loca-

tion 350 and transfers program control
to location 351.)

201 DCA 270 (This instruction stores the contents of

the AC in location 270 upon return
from the subroutine.)

2-12

SUBROUTINE

350 0000 (This location is assumed to have an

initial value of 0000; after JMS 350 is
executed, it is 0201.)

351 iii (First instruction of subroutine)

375 JMP I 350 (Last instruction of subroutine)

The following should be kept in mind when using the JMS:

1. The value of the PC (the address of the JMS instruction +1)
is always stored in the first location of the subroutine, replacing
the original contents;

2. Program control is always transferred to the location designated
by the operand +1 (second location of the subroutine);

3. The normal return from a subroutine is made by using an in-
direct JMP to the first location of the subroutine (JMP I 350
in the above example); (Indirect addressing, as discussed later
in this chapter, effectively transfers control to location 201.);

4. When the results of the subroutine processing are contained in
the AC and are to be used in the main program, they must be
stored upon return from the subroutine before further calcula-
tions are performed. (In the above example, the results of the
subroutine processing are stored in location 270. )

ADDRESSING

When the memory reference instructions were introduced, it was
stated that nine bits are allocated to specify the operand (the address
referenced by the instruction). The method used to reference a memory
location using these nine bits will now be discussed.

PDP-8 Memory Pages

As previously described, the. format of an MRI is three bits (0, 1,
and 2) for the operation code and the remaining nine bits the operand.
However, a full twelve bits are needed to uniquely address the 4,096
(10,000 octal) locations that are contained in the PDP-8 memory unit.
To make the best use of the available nine bits, the PDP-8 utilizes a
logical division of memory into blocks (pages) of 200 8 locations each,
as shown in the following table.

2-13

Page

Memory
Locations

Page

Memory
Locations

0-177

4000-4177

200-377

4200-4377

400-577

4400-4577

600-777

4600-4777

1000-1177

5000-5177

1200-1377

5200-5377

1400-1577

5400-5577

1600-1777

5600-5777

2000-2177

6000-6177

2200-2377

6200-6377

2400-2577

6400-6577

2600-2777

6600-6777

3000-3177

7000-7177

3200-3377

7200-7377

3400-3577

7400-7577

3600-3777

7600-7777

Since there are 200 8 locations on a page and seven bits can represent
200 8 different numbers, seven bits (5 through 11 of the MRI) are used
to specify the page address. Before discussing the use of the page ad-
dressing convention by an MRI, it should be emphasized that memory
does not contain any physical page separations. The computer recog-
nizes only absolute addresses and does not know what page it is on, or
when it enters a different page. But, as will be seen, page addressing
allows the programmer to reference all of the 4,096 10 locations of
memory using only the nine available bits of an MRI. The format of an
MRI is shown in Figure 2-3.

OPERAND

BIT
POSITION O

EACH BIT IS

EITHER I
OR I

OPERATION
CODE

ADDRESS MODE BIT

O: DIRECT ADDRESSING

I : INDIRECT ADDRESSING

PAGE ADDRESS BITS
(O TO I77 8 )

CURRENT PAGE OR PAGE O BIT
o: PAGE
I : CURRENT PAGE

Figure 2-3. Format of a Memory Reference Instruction
2-14

As previously stated, bits through 2 are the operation code for the
MRI. Bits 5 through 11 identify a specific location on a given page, but
they do not identify the page itself. The page is specified by bit 4, often
called the current page or page bit. If bit 4 is a 0, the page address is
interpreted as a location on page 0. If bit 4 is a 1, the page address
specified is interpreted to be on the current page (the page on which
the MRI itself is stored). For example, if bits 5 through 11 represent
123 8 and bit 4 is a 0, the location referenced is absolute address 123 R .
However, if bit 4 is a 1 and the current instruction is in a core memory
location whose absolute address is between 4,600 8 and 4,777 8 , the page
address 123 s designates the absolute address 4,723 g . Note that, as
shown in the following example, this characteristic of page addressing
results in the octal coding for two TAD instructions on different
memory pages being identical when their operands reference the same
relative location (page address) on their respective pages.

Content

Location

Mnemonic

Octal

Explanation

200
400

TAD 250
TAD 450

1250
1250

TAD 250 and TAD 450 both
mean add the contents of loca-
tion 50 on the current page (bit
4 = 1 ) to the accumulator.

Except when it is on page 0, a memory reference instruction can refer-
ence 400 8 locations directly, namely those 200 8 locations on the page
containing the instruction itself and the 200 8 locations on page 0, which
can be addressed from any memory location.

NOTE: If an MRI is stored in one of the first 200 8 memory locations (0 to
177 8 ), current page is page 0; therefore, only locations to 177 8 are
directly addressable.

Indirect Addressing

In the preceding section, the method of directly addressing 400 s
memory locations by an MRI was described — namely those on page
and those on the current page. This section describes the method for
addressing the other 7400 8 memory locations. Bit 3 of an MRI, shown
in Figure 2-3 but not discussed in the preceding section, designates the
address mode. When bit 3 is a 0, the operand is a direct address. When
bit 3 is a 1, the operand is an indirect address. An indirect address
(pointer address) identifies the location that contains the desired address
(effective address). To address a location that is not directly address-
able, the absolute address of the desired location is stored in one of
the 400 8 directly addressable locations (pointer address); the pointer
address is written as the operand of the MRI; and the letter I is written

2-15

between the mnemonic and the operand. (During assembly, the pres-
ence of the I results in bit 3 of the MRr being set to 1 .) Upon execution,
the MRI will operate on the contents of the location identified by the
address contained in the pointer location.

The two examples in Figure 2-4 illustrate the difference between
direct addressing and indirect addressing. The first example shows a
TAD instruction that uses direct addressing to get data stored on page
in location 50; the second is a TAD instruction that uses indirect ad-
dressing, with a pointer on page in location 50, to obtain data stored
in location 1275. (When references are made to them from various
pages, constants and pointer addresses can be stored on page to avoid
the necessity of storing them on each applicable page.) The octal value
1050, in the first example, represents direct addressing (bit 3 = 0); the
octal value 1450, in the second example, represents indirect addressing
(bit 3 = 1). Both examples assume that the accumulator has previously
been cleared.

Location Content

200

TAD

1275

(TAD 50 = 1050.)

Address

Instruction

■Data (Number) To Be Acted Upon By
■Instruction Address
(Content of location 1275 is not used in
the execution of the instruction in loca-
tion 200.)
NOTE: AC = 1275 after executing the instruction in loca-
tion 200.

Location Content

(TAD I 50 = 1450 s )
Pointer Address
Designates Indirect Addressing
Instruction

•Effective Address
Pointer Address

-Data (Number) To Be Acted Upon By
Instruction
Effective Address
NOTE: AC = 20 after executing the instruction in location
200-

Figure 2-4. Comparison of Direct and Indirect Addressing

2-16

The following three examples illustrate some additional ways in
which indirect addressing can be used. As shown in example 1, indirect
addressing makes it possible to transfer program control off page (to
any desired memory location). (Similarly, indirect addressing makes it
possible for other memory reference instructions to address any of the
4,096k, memory locations.) Example 2 shows a DCA instruction that
uses indirect addressing with a pointer on the current page. The pointer
in this case designates a location off the current page (location 227) in
which the data is to be stored. (A pointer address is normally stored on
the current page when all references to the designated location are from
the current page.) Indirect addressing provides the means for returning
to a main program from a subroutine, as shown in example 3. Indirect
addressing is also effectively used in manipulating tables of data as de-
scribed and illustrated in conjunction with autoindexing in Chapter 3.

EXAMPLE 1
Location

100

Content
JMP I 100.

6000

(JMP 1 100 = 5500 8 )
-Pointer Address

Designates Indirect Addressing
- Instruction

NOTE:

■ Effective Address

Pointer Address
DCA6J00

Next Instruction To Be Executed

Effective Address
Execution of the instruction in location 75 causes,, pro-
gram control to be transferred to location 6000, and
the next instruction to be executed is the DCA 6100
instruction.

EXAMPLE 2
Location

Content

(DCA I 577 = 3777 8 )
Pointer Address
Designates Indirect Addressing
Instruction

NOTE:

• Effective Address

■ Pointer Address
nnnn.,

-Data (Number) Stored By Instruction

Effective Address
Execution of the, instruction in location 450 causes the
contents of the accumulator to be stored in location
227.

2-17

EXAMPLE 3
Location

207
210

2000
2001

Content

JMS 1 70
TAD 250

2000

aaaa

(JMS I 70 = 4470»)

(The next instruction to be executed

upon return from the subroutine.)

(Starting address of the subroutine
stored here.)

(Return address stored here by JMS

instruction.)

(First instruction of subroutine.)

2077

JMP I 2000

(Last instruction of subroutine.)

NOTES: I. Execution of the instruction in location 207 causes
the address 210 to be stored in location 2000 and
the instruction in location 2001 to be executed
next. Execution of the subroutine proceeds until
the last instruction (JMP I 2000) causes control
to be transferred back to the main program, con-
tinuing with the execution of the instruction stored
in location 210.

2. A JMS instruction that uses indirect addressing is
useful when the subroutine is too large to store on
the current page.

3. Storing the pointer address on page enables in-
structions on various pages to have access to the
subroutine.

OPERATE MICROINSTRUCTIONS

The operate instructions (octal operation code = 7) allow the pro-
grammer to manipulate and/or test the data that is located in the
accumulator and link bit. A large number of different instructions are
possible with one operation code because the operand bits are not
needed to specify an address as they are in an MRI and can be used to
specify different instructions. The operate instructions are separated
into two groups: Group 1, which contains manipulation instructions,
and Group 2, which is primarily concerned with testing operations.
Group 1 instructions are discussed first.

Group 1 Microinstructions

The Group 1 microinstructions manipulate the contents of the accu-
mulator and link. These instructions are microprogrammable; that is,
they can be combined to perform specialized operations with other
Group 1 instructions. Microprogramming is discussed later in this
chapter.

2-18

10 II

CLA CLL CMACML

RAR

RAL

I AC

OPERATION
CODE

ZERO SPECIFIES
GROUP I

: ROTATE ONE PLACE
I : ROTATE TWO PLACES

The preceding diagram illustrates the manner in which a PDP-8 in-
struction word is interpreted when it is used to represent a Group 1
operate microinstruction. As previously mentioned, 7 8 is the operation
code for operate microinstructions; therefore, bits through 2 are all
l's. Since a reference to core memory is not necessary for the operation
of microinstructions, bits 3 through 11 are not used to reference an
address. Bit 3 contains a to signify that this is a Group 1 instruction,
and the remaining bits are used to specify the operations to be per-
formed by the instruction. The operation of each individual instruction
specified by these bits is described below.

CLA

\CLL
CMA

CML
RAR

Clear the accumulator. If bit 4 is a 1, the instruction sets
the accumulator to all zeroes.
Clear the link. If bit 5 is a 1, the link bit is set.to 0.
Complement the accumulator. If bit 6 is a 1, the accumu-
lator is set to the l's complement of its original value; that
is, all l's become O's, and all O's become l's.
Complement the link. If bit 7 is a 1, the state of the link bit
is reversed.

Rotate the accumulator and link right. If bit 8 is a 1 and
bit 10 is a 0, the instruction treats the AC and L as a closed
loop and shifts all bits in the loop one position to the right.
This operation is illustrated by the following diagram.

00011 ioooooo- 1

AWWWWW

IOOOII 100000

BEFORE RAR

AFTER RAR

RTR

Rotate the accumulator and link twice right. If bit 8 is a 1
and bit 10 is also a 1, a shift of two places to the right is
executed. Both the RAR and RTR instructions use what is
commonly called a circular shift, meaning that any bit
rotated off one end of the accumulator will reappear at the
other end. This operation is illustrated below.

2-19

1££s

BEFORE RTR

1 1 10

RAL Rotate the accumulator and link left. If bit 9 is a 1 and bit
10 is a 0, this instruction treats the AC and L as a closed
loop and shifts all bits in the loop one position to the left,
performing a circular shift to the left.

RTL Rotate the accumulator and link twice left. If bit 9 is a 1
and bit 10 is a 1 also, the instruction rotates each bit two
positions to the left. (The RAL and RTL microinstructions
shift the bits in the reverse direction of that directed by the
RAR and RTR microinstructions. )

IAC Increment the accumulator. When bit 11 is a 1, the con-

tents of the AC is increased by 1 .

NOP No operation. If bits through 2 contain operation code
7 S , and the remaining bits contain zeros, no operation is
performed and program control is transferred to the next
instruction in sequence.

A summary of Group 1 instructions, including their octal forms, is
given below.

Mnemoni

c 1 Octal 2

NOP

7000

CLA

7200

CLL

7100

CMA

7040

CML

7020

RAR

7010

RAL

7004

RTR

7012

RTL

7006

IAC

7001

Operation

No operation

Clear AC

Clear link bit

Complement AC

Complement link bit

Rotate AC and L right one position

Rotate AC and L left one position

Rotate AC and L right two positions

Rotate AC and L left two positions

Increment AC

Sequence 8

1
1
2
2
4
4
4
4
3

1 Mnemonic code is meaningful to and translated by an assembler into binary

code.

2 Octal numbers conveniently represent binary instructions.

3 Sequence numbers indicate the order in which the operations are performed

by the PDP-8/I and PDP-8/L (sequence 1 operations are performed first,
sequence 2 operations are performed next, etc.).

2-20

Group 2 Microinstructions

Group 2 operate microinstructions are often referred to as the "skip
microinstructions" because they enable the programmer to perform
tests on the accumulator and link and to skip the next instruction de-
pending upon the results of the test. They are usually followed in a pro-
gram by a JMP (or possibly a JMS) instruction. A skip instruction
causes the computer to check for a specific condition, and, if it is pres-
ent, to skip the next instruction. If the condition were not present, the
next instruction would be executed.

TO SPECIFY
GROUP 2

CLA

SMA/

/spa

SZA/
/SNA

SNL/
/SZL

0/1/
/-SKP

OSR

HLT

OPERATION

VALUE OF BIT 8

CODE 7„

DETERMINES THE

CONTAINS A J

NITAIlN

S A

ACTION SPECIFIED
BY BITS 5,6,5 7

TO SPECIFY

GROUP 2

REVERSE SENSING BIT
0: SMA.SZA.aSNL ARE ENABLED.
I : SPA, SNA.aSZL ARE ENABLED.
(UNCONDITIONAL SKIP WHEN
BITS 5,6,8 7 ARE 0"S)

The available instructions are selected by bit assignment as shown in
the above diagram. The operation of each individual instruction speci-
fied by these bits is described below.

CLA Clear the accumulator. If bit 4 is a 1, the instruction sets

the accumulator to all zeros.
SMA Skip on minus accumulator, li bit 5 is a 1 and bit 8 is au,

the next instruction is skipped if the accumulator is less

than zero.
SPA Skip on positive accumulator. If bit 5 is a ,1 and bit 8 is a

1, the next instruction is skipped if the accumulator is

greater than or equal to zero.
SZA skip on nonzero accumulator. If bit 6 is a 1 and bit 8 is a

1, the next instruction is skipped if the accumulator is not

zero.
SNA Skip on nonzero accumulator. If bit 6 is a 1 and bit 8 is a

1 also, the next instruction is skipped if the accumulator is

not zero.

2-21

SNL Skip on nonzero link. If bit 7 is a 1 and bit 8 is a 0, the
next instruction is skipped when the link bit is a 1.

SZL Skip on zero link. If bit 7 is a 1 and bit 8 is a 1, the next

instruction is skipped when the link bit is a 0.

SKP Unconditional skip. If bit 8 is a 1 and bit 5, 6 and 7 are

all zeros, the next instruction is skipped. (Bit 8 is a reverse
sensing bit when bits 5, 6 or 7 are used — see SMA, SPA,
SZA, SNA, SNL, and SZL above. )

OSR Inclusive OR of switch register with AC. If bit 9 is a 1, an
inclusive OR operation is performed between the content
of the accumulator and the console switch register. The re-
sult is left in the accumulator and the original content of
the accumulator is destroyed. In short, the inclusive OR
operation consists of the comparison of the corresponding
bit positions of the two numbers and the insertion of a 1 in
the result if a 1 appears in the corresponding bit position
in either number. See Chapter 1 for further discussion. The
action of the instruction is illustrated below.

.'.

ACCUMULATOR
SWITCH REGISTER

1 1 I

RESULT IN AC

HLT Halt. If bit 10 is a 1, the computer will stop at the conclu-
sion of the current machine cycle.

A summary of Group 2 instructions, including their octal representa-
tion, is given in the following table.

Mnemonic

Octal

Operation

Sequence

CLA

7600

Clear the accumulator

SMA

7500

Skip on minus accumulator

SPA

7510

Skip on positive accumulator
(or AC = 0)

SZA

7440

Skip on zero accumulator

SNA

7450

Skip on nonzero accumulator

SNL

7420

Skip on nonzero link

SZL

7430

Skip on zero link

SKP

74 10

Skip unconditionally

OSR

7404

Inclusive OR, switch register
with AC

HLT

7402

Halts the program

2-22

MICROPROGRAMMING

Because PDP-8 instructions of Group 1 and Group 2 are determined
by bit assignment, these instructions may be combined, or micropro-
grammed, to form new instructions enabling the computer to do more
operations in less time.

Combining Microinstructions

The programmer should make certain that the program clears the
accumulator and link before any arithmetic operations are performed.
To perform this task, the program might include the following instruc-
tions (given in both octal and mnemonic form).

CLA
CLL

7200 (octal)
7100 (octal)

However, when the Group 1 instruction format is analyzed, the follow-
ing is observed.

CLA

.CLL

OPERATION
CODE

MUST BE A I TO SPECIFY CLL
MUST BE A I TO SPECIFY CLA

MUST BE A TO SPECIFY GROUP I

Since the CLA and the CLL instructions occupy separate bit posi-
tions, they may be expressed in the same instruction, thus combining
the two operations into one instruction. This instruction would be writ-
ten as follows.

CLA CLL

7300 (octal)

In this manner, many operate microinstructions can be combined mak-
ing the execution of the program much more efficient. The assembler
for the PDP-8 will combine the instructions properly when they are
written as above, that is, on the same coding line, and separated by a
space.

Illegal Combinations

Microprogramming, although very efficient, can also be troublesome
for the new programmer. There are many violations of coding which
the assembler will not accept.

2-23

One rule to remember is : "If you can't code it, the computer can't do
it." In other words, the programmer could write a string of mnemonic
microinstructions, but unless these microinstructions can be coded cor-
rectly in octal representation, they cannot be performed. To illustrate
this fact, suppose the programmer would like to complement the accu-
mulator (CMA), complement the link (CML), and then skip on a
nonzero link (SNL). He could write the following.

CMA CML SNL
These instructions require the following bit assignments.

1 23456789 10 II

CMA

CML

SNL I I I I

The three microinstructions cannot be combined in one instruction be-
cause bit 3 is required to be a and a 1 simultaneously. Therefore, no
instructions may be used which combine Group 1 and Group 2 micro-
instructions because bit 3 usage is not compatible. The CMA and CML
can, however, be combined because their bit assignments are com-
patible. The combination would be as follows.

CMA CML 7060 (octal)

To perform the original set of three operations, two instructions are
needed.

CMA CML
SNL

7060 (octal)
7420 (octal)

Because Group 1 and Group 2 microinstructions cannot be com-
bined, the commonly used microinstruction CLA is a member of both
groups. Clearing the AC is often required in a program and it is very
convenient to be able to microprogram the CLA with the members of
both groups.

The problem of bit assignment also arises when some instructions
within a group are combined. For example, in Group 1 the rotate in-
structions specify the number of places to be rotated by the state of bit
10. If bit 10 is a 0, rotate one place; if bit 10 is a 1, rotate two places.
Thus, the instruction RAL can not be combined with RTL because bit
10 would be required to have two different values at once. If the pro-

2-24

grammer wishes to rotate right three places, he must use two separate
instructions.

RAR 7010 (octal)

RTR 7012 (octal)

Although he can write the instruction "RAR RTR", it cannot be cor-
rectly converted to octal by the assembler because of the conflict in bit
10; therefore, it is illegal.

Combining Skip Microinstructions

Group 2 operate microinstructions use bit 8 to determine the instruc-
tion specified by bits 5, 6, and 7 as previously described. If bit 8 is a 0,
the instructions SMA, SZA, and SNL are specified. If bit 8 is a 1, the
instructions SPA, SNA, and SZL are specified. Thus, SMA cannot be
combined with SZL because of the opposite values of bit 8. The skip
condition for combined microinstructions is established by the skip con-
ditions of the individual mcroinstructions in accordance with the rules
for logic operations (see "Logic Primer" in Chapter 1).
OR GROUP— SMA OR SZA OR SNL

If bit 8 is a 0, the instruction skips on the logical OR of the condi-
tions specified by the separate microinstructions. The next instruction
is skipped if any of the stated conditions exist. For example, the com-
bined microinstruction SMA SNL wUl skip under the following condi-
tions:

1. The accumulator is negative, the link is zero.

2. The link is nonzero, the accumulator is not negative.

3. The accumulator is negative and the link is nonzero.

(It will not skip if all conditions fail.) This manner of combining the
test conditions is described as the logical OR of the conditions.
AND GROUP— SPA AND SNA AND SZL

A value of bit 8 = 1 specifies the group of microinstructions SPA,
SNA, and SZL which combine to form instructions which act according
to the logical AND of the conditions. In other words, the next instruc-
tion is skipped only if all conditions are satisfied. For example, the in-
struction SPA SZL will cause a skip of the next instruction only if the
accumulator is positive and the link is zero. (It will not skip if either
of the conditions fail. )

NOTES? 1. The programmer is not able to specify the manner
of combination. The SMA, SZA, SNL conditions
are always combined by the logical OR, and the
SPA, SNA, SZL conditions are always joined by a
logical AND.
2. Since the SPA microinstruction will skip on either
a positive or a zero accumulator, to skip on a
strictly positive (positive, nonzero) accumulator
the combined microinstruction SPA SNA is used.

2-25

Order of Execution of Combined Microinstructions

The combined microinstructions are performed by the computer in a
very definite sequence. When written separately, the order of execution
of the instructions is the order in which they are encountered in the pro-
gram. In writing a combined instruction of Group 1 or Group 2 micro-
instructions, the order written has no bearing upon the order of
execution. This should be clear, because the combined instruction is a
12-bit binary number with certain bits set to a value of 1. The order in
which the bits are set to 1 has no bearing on the final execution of the
whole binary word.

The definite sequence, however, varies between members of the
PDP-8 computer family. The sequence given here applies to the PDP-
8/1 and PDP-8/L. The applicable information for other members of
the PDP-8 family is given in Appendix E. The order of execution for
PDP-8/I and PDP-8/L microinstructions is as follows.

GROUP 1

Event 1

Event 2
Event 3

Event 4

CLA, CLL — Clear the accumulator and/or clear the
link are the first actions performed. They are effectively
performed simultaneously and yet independently.
CMA, CML — Complement the accumulator and/or com-
plement the link. These operations are also effectively
performed simultaneously, and independently.
IAC — Increment the accumulator. This operation is per-
formed third allowing a number in the AC to be comple-
mented and then incremented by 1, thereby forming the
two's complement, or negative, of the number.
RAR, RAL, RTR, RTL — The rotate instructions are per-
formed last in sequence. Because of the bit assignment
previously discussed, only one of the four operations may
be performed in each combined instruction.

GROUP 2
Event 1

Either SMA or SZA or SNL when bit 8 is a 0. Both SPA
and SNA and SZL when bit 8 is a 1. Combined micro-
instructions specifying a skip are performed first. The
microinstructions are combined to form one specific test,
therefore, skip instructions are effectively performed
simultaneously.

Because of bit 8, only members of one skip group may be
combined in an instruction.

2-26

Event 2 CLA — Clear the accumulator. This instruction is per-
formed second in sequence thus allowing different arith-
metic operations to be performed after testing (see Event
1 ) without the necessity of clearing the accumulator with
a separate instruction before some subsequent arithmetic
operation.

Event 3 OSR — Inclusive OR between the switch register and the
AC. This instruction is performed third in sequence,
allowing the AC to be clear,ed first, and then loaded from
the switch register.

Event 4 HLT — The HLT is performed last to allow any other
operations to be concluded before the program stops.

This is the order in which all combined instructions are performed.
In order to perform operations in a different order, the instructions
must be written separately as shown in the following example. One
might think that the following combined microinstruction would clear
the accumulator, perform an inclusive OR between the SR and the AC,
and then skip on a nonzero accumulator.

CLA OSR SNA
However, the instruction would not perform in that proper manner,
because the SNA would be executed first. In order to perform the skip
last, the instructions must be separated as follows.

CLA OSR

SNA

Microprogramming requires that the programmer carefully code
mnemonics legally so that the instruction does in fact do what he desires
it to do. The sequence in which the operations are performed and the
legality of combinations is crucial to PDP-8 programming.

The following is a list of commonly used combined microinstructions,
some of which have been assigned a separate mnemonic.

Instruction Explanation

— CLA CLL Clear the accumulator and link.

CIA CMA IAC Complement and increment the accumulator.

(Sets the accumulator equal to its own nega-
tive.)
LAS CLA OSR Load accumulator from switches.

(Loads the accumulator with the value of the
switch register.)
STL CLL CML Set the link (to a 1 ) .

— CLA IAC Sets the. accumulator to a 1 .

STA

CLA CMA Sets the accumulator to a — 1.

2-27

In summary, the basic rules for combining operate microinstructions
are given below.

1. Group 1 and Group 2 microinstructions cannot be combined.

2. Rotate microinstructions (Group 1) cannot be combined with
each other.

3. OR Group (SMA, SZA, or SNL) microinstructions cannot be
combined with AND Group (SPA, SNA, or SZL) microin-
structions.

4. OR Group microinstructions are combined as the logical OR
of their respective skip conditions. AND Group microinstruc-
tions are combined as the logical AND of their respective skip
conditions.

5. Order of execution for combined instructions (PDP-8/I and
PDP-8/L only) is listed below.

Group 1 Group 2

CLA, CLL

SMA/ SZA/ SNL or
SPA/SNA/SZL

CMA, CML

CLA

IAC

OSR

RAR, RAL, RTR, RTL

HLT

EXERCISES

1. The following is a list of current addresses and locations to be
addressed. Determine whether the second location should be di-
rectly or indirectly addressed from the first.

Currer

>t Address
2456

Location to be Addressed

2577

1500

1600

1230

0030

0050

0120

6555

6400

6555

6600

4343

4100

2742

2450

2507

5507

J-

3200

3377

2-28

2. What type of instruction is each of the following (MRI, operate
Group 1 or operate Group 2 microinstruction)?

7430

0024

7240

7000

4706

7700

3. Why are each of the following not legal instructions for the PDP-8?
a. 6509 b. 15007 c. 1581 d. 635 e. 7778

4. What is the effect of each of the following octal instructions?

Octal Mnemonic Operation

a. 0000

b. 4010

c. 2300

d. 1777

e. 3500

f. 5400

g. 1030
h. 2577
i. 5273
j. 3150

5. Separate the following octal instructions into microinstruction
mnemonics.

a. 7260

b. 7112

c. 7440

d. 7632

e. 7550

f. 7007
f. 7770

6. Write the octal representation for each location in the following
program. What are the contents of the accumulator and locations
205, 206, and 207 after execution of the program?

Octal

Location

Mnemonic

0200

CLA

0201

TAD 0205

0202

TAD 0206

0203

DCA 0207

0204

HLT

0205

1537

0206

2241

0207

0000

2-29

7. Write the octal form of the following microinstructions. Identify
any illegal combinations.

CLA CLL CM A CML

CLL RTL HLT

SPA CLA

CLA IAC RTL

CLA IAC RAL RTL

SMA SZA CLA

SMA SZL

CLA OSR HLT

CLA OSR IAC

]•

CLA SMA SZA

8. What instructions could be used to perform a skip only if the
accumulator is zero and the link is nonzero?

9. Why is it not possible to write one combined microinstruction that
will load the accumulator from the console switch register, and
then test that number, skipping on a positive value?

10. Write the following programs.

a. Program starts in location 0200 and adds 2 and 8. Give both
mnemonic and octal representations.

b. Program beginning in location 400 which interchanges the con-
tents of locations 550 and 551. Give both mnemonic and octal
representations.

11. Write programs to add three numbers A, B, and C in the specified
locations below and put the result in the given address for the
SUM. All programs start in location 200. Give octal and mnemonic
coding.

SUM

0030

0031

0032

0033

0300

0301

0302

0303

3000

3001

3002

3003

2-30

Chapter 3

Elementary

Programming

Techniques

Mastery of the instruction set is the first step in learning to program
the PDP-8 family computers. The next step is to learn to use the in-
struction set to obtain correct results and to obtain them efficiently. This
is best done by studying the following programming techniques. Exam-
ples, which should further familiarize the reader with the instructions
and their uses, are given to illustrate each technique.

The modern digital computer is capable of storing information, per-
forming calculations, making decisions based on the results and arriving
at a final solution to a given problem. The computer cannot, however,
perform these tasks without direction. Each step which the computer is
to perform must first be worked out by the programmer.

The programmer must write a program, which is a list of instructions
for the computer to follow to arrive at a solution for a given problem.
This list of instructions is based on a computational method, sometimes
called an algorithm, to solve the problem. The list of instructions is
placed in the computer memory to activate the applicable circuitry so
that the computer can process the problem. This chapter describes the
procedure to be followed when writing a program to be used on the
PDP-8 family of computers.

3-1

PROGRAMMING PHASES

In order to successfully solve a problem with a computer, the pro-
grammer proceeds through the five programming phases listed below:

1. Definition of the problem to be solved,

2. Determination of the most feasible solution method,

3. Design and analysis of the solution- — flowcharting,

A. Coding the solution in the programming language, and
5. Program checkout.

The definition of the problem is not always obvious. A great amount
of time and energy can be wasted if the problem is not adequately de-
fined. When the problem is to sum four numbers, the defining phase is
obvious. However, when the problem is to monitor and control a per-
formance test for semiconductors, a precise definition of the problem
is necessary. The question that must be answered in this phase is:
"What precisely is the program to accomplish?"

Determining the method to be followed is the second important
phase in solving a problem with a computer. There are perhaps an in-
finite number of methods to solve a problem, and the selection of one
method over another is often influenced by the computer system to be
used. Having decided upon a method based on the definition of the
problem and the capabilities of the computer system, the programmer
must develop the method into a workable solution.

The programmer must design and analyze the solution by identifying
the necessary steps to solve the problems and arranging them in a
logical order, thus implementing the method. Flowcharting is a graphical
means of representing the logical steps of the solution. The flowcharting
technique is effective in providing an overview of the logical flow of a
solution, thereby enabling further analysis and evaluation of alternative
approaches.

Having designed the problem solution, the programmer begins coding
the solution in the programming language. This phase is commonly
called programming but is actually coding and is only one part of the
programming process. When the program has been coded and the pro-
gram instructions have been stored in the computer memory, the prob-
lem can be solved. At this point, however, the programming process
is rarely complete. There are very few programs written which initially
function as expected. Whenever the program does not work properly,
the programmer is forced to begin the fifth step of programming, that
of checking out or "debugging" the program.

3-2

The program checkout phase requires the programmer to methodi-
cally retrace the flow of the instructions step-by-step to find any pro-
gram errors that may exist. The programmer cannot tell a computer:
"You know what I mean!", as he might say in daily life. The computer
does not know what is meant until it is told, and once given a set of
instructions, the computer follows them precisely. If needed instructions
are left out or coding is done incorrectly, the results may be surprising.
These flaws, or "bugs" as they are often called, must be found and
corrected. There are many different approaches to finding bugs in a
program; however, the chosen approach must be organized and pains-
takingly methodical if it is to be successful. Several techniques for de-
bugging programs for the PDP-8 family of computers are described
in Chapter 6.

FLOWCHARTING

A simple problem to add three numbers together is solved in a few,
easily determined steps. A programmer could sit at his desk and write
out three or four instructions for the computer to solve the problem.
However, he probably could have added the same three numbers with
paper and pencil in much less time than it took him to write the pro-
gram. Thus, the problems which the programmer is usually asked to
solve are much more complex than the addition of three numbers, be-
cause the value of the computer is in the solution of problems which are
inconvenient or time consuming by human standards.

When a more complex problem is to be solved by a computer, the
program involves many steps, and writing it often becomes long and
confusing. A method for solving a problem which is written in words
and mathematical equations is extremely hard to follow, and coding
computer instructions from such a document would be equally difficult.
A technique called flowcharting is used to simplify the writing of pro-
grams. A flowchart is a graphical representation of a given problem,
indicating the logical sequence of operations that the computer is to
perform. Having a diagram of the logical flow of a program is a tre-
mendous advantage to the programmer when he is determining the
method to be used for solving a problem, as well as when he writes
the coded program instructions. In addition, the flowchart is often a
valuable aid when the programmer checks the written program for
errors.

3-3

The flowchart is basically a collection of boxes and lines. The boxes
indicate what is to be done and the lines indicate the sequence of the
boxes. The boxes are of various shapes which represent the action to
be performed in the program. Appendix C is a guide to the flowchart
symbols and procedures which are used in this text.

The following are examples of flowcharts for specific problems, illus-
trating methods of attacking problems with a computer program as well
as illustrating flowcharting techniques. Example 1 adds three numbers
together. Example 2 puts three numbers in increasing order.

Example 1 — Straight-Line Programming

Example 1 is an illustration of straight-line programming. As the
flowchart shows, there is a straight-line progression through the process-
ing steps with no change in course. The value of X, which is equal to
A+B-j-C is in the accumulator when the program stops.

START

CLEAR
ACCUMULATOR

GET A INTO
ACCUMULATOR

ADD B

ADD C

STOP

Example 1 — Add Three Numbers

Example 2 — Program Branching

Example 2 is designed to arrange three numbers in increasing order.
The program must branch to interchange numbers that are out of
order. (Branching, a common feature of programming, is described in
detail later in this chapter. ) Note that the arithmetic operations of sub-
traction are done in the accumulator, which must be cleared initially.

3-4

START

CLEAR THE
ACCUMULATOR

GET FIRST
NUMBER INTO
ACCUMULATOR

SUBTRACT
SECOND NUMBER

j^ \$ AC ^\

VES

JNO

INTERCHANGE
1ST AND 2ND

NUMBERS

COMPARE
2ND AND 3RD

NUMBERS
AS ABOVE

IS AC

YES

. POSITIVE .
|N0

INTERCHANGE

2ND AND 3RD

NUMBERS

COMPARE
1ST AND 2ND

NUMBERS
AS ABOVE

^ IS AC ^

POSITIVE

YES

INTERCHANGE

1ST AND 2ND

NUMBERS

DONE

)

Example 2 — Arrange Three Numbers in Increasing Order

3-5

CODING A PROGRAM

The introduction of an assembler in Chapter 2 enabled the pro-
grammer to write a symbolic program using meaningful mnemonic codes
rather than the octal representation of the instructions. The programmer
could now write mnemonic programs such as the following example,
which multiplies 18i by 36i using successive addition. '

200/

CLA CLL

(Initialize)

201/

TAD 210

(Set up a Tally

202/

CIA

equal to — 18 10 to

203/

DCA212

count the additions of 36)

204/

TAD 211

(Add 36)

205/

ISZ212

(Skip if Tally is 0)

206/

JMP 204

(Add another 36 if not done)

207/

HLT

(Stop after 18 times)

210/

0022

(Equal to 18 ltl )

211/

0044

(Equal to 36 10 )

212/

0000

(Holds the tally)

Writing the above program was greatly simplified because mnemonic
codes were used for the octal instructions. However, writing down the
absolute address of each instruction is clearly an inconvenience. If the
programmer later adds or deletes instructions, thus altering the location
assignments of his program, he has to rewrite those instructions whose
operands refer to the altered assignments. If the programmer wishes to
move the program to a different section of memory, he must rewrite the
program. Since such changes must be made often, especially in large
programs, a better means of assigning locations is needed. The assem-
bler provides this better means.

Location Assignment

As in the previous program example, most programs are written in
successive memory locations. If the programmer assigned an absolute
location to the first instruction, the assembler could be told to assign
the next instructions to the following locations in order. In programming
the PDP-8, the initial location is denoted by a precedent asterisk (*).
The assembler maintains a current location counter by which it assigns
successive locations to instructions. The asterisk causes the current
location counter to be set to the value following the asterisk. With this
improvement incorporated, the previous example appears as shown
in the following example.

3-6

*200

CLA CLL
TAD 210
CIA

DCA 212
TAD 211
ISZ 212
JMP 204
HLT
0022
0044
0000

NOTE: In this example, CLA CLL is stored in location, 200 and the
successive instructions are stored in 201, 202, etc.

Symbolic Addresses

The programmer does not at the outset know which locations he will
use to store constants or the tally. Therefore he must leave blanks after
each MRI and come back to fill these in after he has assigned locations
to these numbers. In the previous program, he must count the number
of locations after the assigned initial address in order to assign the
correct values to the MRI operands. Actually this is not necessary, be-
cause he may assign symbolic names (a symbol followed by a comma
is a symbolic address) to the locations to which he must refer, and the
assembler will assign address values for him. The assembler maintains
a symbol table in which it records the octal values of all symbolic
addresses. With symbolic address name tags, the program is as shown
below.

*200

START, CLA CLL

TAD A

CIA

DCA TALLY
MULT, TAD B

ISZ TALLY

IMP MULT

HLT

A, 0022

B, 0044
TALLY, 0000

NOTES: 1. The dollar sign is the terminal character for the assembler.

2. The comma after a symbol (e.g., START,) indicates to the
assembler that the Symbol is a symbolic address.

3-7

Symbolic Programming Conventions

Any sequence of letters (A, B, C . . . , Z) and digits (0, 1 , . . . , 9)
beginning with a letter and terminated by a delimiting character (see
Table 3-1) is a symbol. For example, the mnemonic codes for the
PDP-8 instructions are symbols for which the assembler retains octal
equivalents in a permanent symbol table.

User-defined symbols (stored in the external symbol table) may be
of any length; however, only the first six characters are considered, and
any additional characters are ignored. (Symbols which are identical in
their first six characters are considered identical.)

Any sequence of digits followed by a delimiting character forms a
number. The assembler will accept numbers which are octal or decimal.
The radix is initially set to octal and remains octal unless otherwise
specified. The pseudo-instruction DECIMAL may be inserted in the
coding to instruct the assembler to interpret all numbers as decimal until
the next occurrence of the pseudo-instruction OCTAL in the coding.
These pseudo-instructions affect all numbers included in the symbolic
program including those preceeded by an * to denote change of origin.

Each symbol or number written in a PDP-8 program must represent
a 12-bit binary value in order to be interpreted by the assembler.

The special characters in Table 3-1 are used to specify operations to
be performed by the assembler upon symbols or numbers in PDP-8
symbolic programs.

The comma after a symbol in a line of coding (e.g., MULT, TAD B)
indicates to the assembler that the value of MULT is the address of
the location in which the instruction is stored. When an instruction that
references MULT (now a symbolic address) is encountered, the assem-
bler supplies the correct address value for MULT. (Care must be taken
that a symbolic address is never used twice in the same program and
that all locations referenced by an MRI are identified somewhere in the
program.)

The space and tab are used to delimit a symbol or number. In a com-
bined microinstruction such as CLA CLL, the space delimits the first
mnemonic from the second, and the assembler combines the two mne-
monic into one instruction. The space and tab similarly delimit the mne-
monic from the symbolic address.

TAD A or TAD A

t-SPACE ^CTRL/TAB

3-8

Table 3-1. Special Characters for the PDP-8 Symbolic Language

Character

Ike

Keyboard

Name

SPACE

space

combine symbols or numbers

(nonprinting)

(delimiting)

CTRL/ TAB

tab (nonprinting)

combine symbols or numbers or for-
mat the symbolic tape (delimiting)

RETURN

carriage return
(nonprinting)

terminate line (delimiting)

plus

combine symbols or numbers

minus

combine symbols or numbers

comma

assign symbolic address

equals

define parameters

asterisk

set current location counter

semicolon

terminate coding line (delimiting)

$' '

dollar sign

terminate pass (delimiting)

point

has value equal to current location
counter

slash

indicates start of a comment

The carriage return is used to terminate a line of coding. The assem-
bler will also recognize a semicolon as a line terminating character.

TAD A
TAD B

is the same as TAD A; TAD B

One of these two characters (i.e., semicolon or carriage return)
must be used to separate each line of coding.

The assembler will recognize the arithmetic symbols -\- and — in
conjunction with numbers or symbols, thereby enabling "address arith-
metic". For example, the instruction JMP START+1 will cause the
computer to execute the instruction in the next location after START.
The numbers specified in such instructions are subject to the pseudo-
instructions DECIMAL and OCTAL, therefore the number is Inter-
preted as an octal number unless the pseudo-instruction DECIMAL is
in effect.

The decimal point, or period, is a character which is interpreted by
the assembler as the value of the current location counter. This special
symbol can be used as the operand of an instruction; for example, the
instruction JMP .—1 causes the computer to execute the preceding in-
struction.

The equal sign is used to define symbols. This character is used to
replace an undefined symbol with the value of a known quantity. For
example, the programmer could define a "new instruction" NEGATE

3-9

by writing that NEGATE =• CIA. The programmer could then write
the following instructions to subtract B from A.

START, TAD B

NEGATE
TAD A
HLT
NEGATE = CIA

The above coding would be assembled as if the instruction CIA had
been included in the actual coding.

The slash is used to insert comments and headings as described later
in this chapter.

The dollar sign as previously noted, is a terminal character for the
assembler itself. When this character is encountered, the assembler
stops accepting input and terminates the assembly pass, as described in
Chapter 6.

These characters and conventions will be used throughout the re-
mainder of this text to code programs in PAL III, the symbolic lan-
guage of the PDP-8 family of computers. Thus, all examples given may
be directly punched on paper tape as described in Chapter 4 and
assembled by the procedure described in Chapter 6.

PROGRAMMING ARITHMETIC OPERATIONS

The instructions for the PDP-8 may be used to perform the basic
arithmetic operations within the limits of the machine to represent the
necessary numbers. That is, numbers may be added unless the sum ex-
ceeds 4095k, or 7777 a . When a sum exceeds the size of the accumulator,
overflow occurs and incorrect answers result. This condition can
usually be detected by checking the value of the link bit.

The following instructions will add numbers and check for overflow,
halting the program if the link is 1.

ADD, CLA CLL

TAD A
TAD B
SZL
HLT
DCA SUM

3-10

Since the link is initially cleared in the above example, a link value
equal to 1 is an indication that the sum of the contents of locations A
and B is too large to be represented by the 12-bit accumulator alone.
The computer will halt if the overflow is detected with the actual sum
in the combined 13 bits of the accumulator and link.

Arithmetic Overflow

Since the PDP-8 regards the numbers through 3777 8 as positive
numbers and the numbers 4000 8 through 7777 8 as negative numbers,
the addition of two positive numbers could result in either a positive
or a negative number depending upon the size of the numbers added.
Arithmetic overflow is said to occur whenever two positive numbers
add to form a negative number, as shown in the following example.

243 3„ ( a positive number )
+22 11 8 (a positive number)
46447 (considered a negative number by the
PDP-8)

Likewise, two negative numbers could be added to yield a positive num-
ber as in the following example.

5275 8 (-2503,)
+576 1„ (-2017,)
Disregarded — >1 3256 (considered a positive number by the

PDP-8)

Because of situations like those illustrated in the two preceding ex-
amples, the programmer must consider the size of the numbers used
in programmed arithmetic operations. If the programmer suspects that
overflow may occur in the result of an arithmetic operation, he should
follow such an operation by a set of instructions to correct the error or
at least to indicate that such an overflow occurred.

The conditions outlined below may be used to test for arithmetic
overflow,

Signs of Numbers Added Overflow and Link Value

Positive + Negative No overflow possible; link value ignored.

Positive + Positive May result in negative sum; no change in

link value.
Negative + Negative May result in positive sum; link is always

complemented regardless of the sign of the

result.

The program coding on the next page uses the following facts, as-
suming an initially cleared link, to quickly determine the sign of the
sum of two unknown quantities, A and B.

3-11

Sign

of A

of B

Positive

Negative

Positive

Negative

Result of Adding only Bit of A to all of J
Link Value Bit of AC

1
1

/CODING TO ADD TWO NUMBERS
/TESTING FOR ARITHMETIC OVERFLOW.

START,

CLA CLL

TAD A

AND MASK

TAD B

SZL

JMP BTHNEG

RAL

SZL CLA

JMP OPPSGN

JMP BTHPOS

OPPSGN,

TAD A

IAD B

DCA SUM

HLT

BTHNEG,

CLA CLL

TAD A

TAD B

SMA

JMP NEGERR

DCA SUM

HLT

BTHPOS,

TAD A

TAD B

SPA

JMP POSERR

DCA SUM

HLT

SUM,

MASK,

4000

nnnn

POSERR,
NEGERR,

/MASK OUT ALL BUT BIT 0.
/ADD B TO BIT OF A.
/LINK = 1 IMPLIES BOTH
/ARE NEGATIVE.
/ROTATE BIT INTO LINK.
/BIT = 1 IMPLIES
/OPPOSITE SIGNS.
/BIT = 0, BOTH POSITIVE.
/IF A AND B ARE OF OPPOSITE
/SIGNS, THE ADDITION
/CANNOT RESULT IN
/OVERFLOW.

/IF TWO NEGATIVE NUMBERS
/ADD TO FORM A
/POSITIVE NUMBER,
/JMP TO ERROR ROUTINE.
/OTHERWISE, STORE SUM.

/IF TWO POSITIVE
/NUMBERS ADD TO FORM
/A NEGATIVE NUMBER, JMP
/TO ERROR ROUTINE.
/OTHERWISE, STORE SUM.

/ANY NUMBERS A AND B

/ROUTINE TO SIGNAL
/ARITHMETIC OVERFLOW
/OF POSITIVE NUMBERS.
/ROUTINE TO SIGNAL
/ARITHMETIC OVERFLOW
/OF NEGATIVE NUMBERS.

3-12

Subtraction

Subtraction in the PDP-8 family of computers is accomplished by
negating the subtrahend (replacing it by its two's complement) and
then adding it to the minuend, ignoring the overflow if any. The follow-
ing example shows the contents of the accumulator for each step of the
subtraction process.

Subtraction Program

Resulting Contents

Link

Accumulator

CLA CLL
TAD B
CMA
IAC
TAD A

000 000 000 000
000 000 011 111
111 111 100 000
111 111 100 001
000 000 000 111

(0000)
(0037)
(7740)
(7741)
(0007)

A,
B,

0046 (000
0037 (000

000
000

100 110)
011 111)

Note that the number to be subtracted (subtrahend) is brought into
the accumulator, complemented (l's complement) and incremented by
1 (to form the 2's complement). (The 2's complement could be ob-
tained directly through the one microinstruction CIA.) The number
from which A is to be subtracted (minuend) is then added to the ac-
cumulator and the difference is obtained.

If A were already in the accumulator from a previous calculation, an
alternate procedure could be followed. The number A could be negated
first, then B added to it to get B-A. Negating this result yields the same
answer because — (B-A) is equal to A-B.

Multiplication and Division

A previous example illustrated the method of performing multiplica-
tion with the basic PDP-8 instructions, namely by repeated addition.
Obviously, multiplication by this method is also subject to the limita-
tion of overflow. The largest positive number which can be directly
represented is 2047 10 or 3777 8 .

Multiplication by repeated addition will properly handle positive and
negative numbers within the limits of positive or negative arithmetic
overflow. For example 7777 8 is the PDP-8 representation for —1. If it
is multiplied by itself the answer should be + 1 . In other words, adding
7777 8 to itself 7777 8 times should leave (after carries from the most
significant bit) the accumulator equal to 1.

3-13

Disregarded
Carries

1st
2nd

3rd

7776th
7777th

0001

Thus, successive addition will work properly as a method of multiply-
ing negative as well as positive numbers in the PDP-8 family of com-
puters.

Similarly, division could be performed by repeated subtraction. This
method of division could be used to obtain a quotient and remainder,
because only whole numbers are directly represented in the PDP-8.
There are, however, much more efficient means of multiplying and di-
viding numbers in the PDP-8. One means is through the extended arith-
metic element (EAE) option, which is described in Chapter 4. Multi-
plication and division can also be performed through use of the floating
point packages, mathematical routines, and interpretive languages of
the system software for the PDP-8. These "software" approaches to
multiplication and division are described in Chapter 6 of this book.

Double Precision Arithmetic

Two memory location (24 bits) are used to express double precision
numbers. Using these 24 bits allows the representation of numbers in
the range — 8X10 6 to 8X10 8 . The following program adds two double
precision numbers, obtaining a double precision result

Note that if the addition of AL and BL produces a carry, it will
appear in the link. The accumulator is cleared by the DCA CL instruc-
tion, and the RAL instruction moves the value of the link into the least
significant bit position. The values of AH and BH are then added to
the carry (if any) and the higher part of the answer is deposited in CH.

3-14

This technique may be extended to any order of multiple precision.

*200

DUBADD,

CLA CLL

TAD AL

TAD BL

DCA CL

RAL

TAD AH

TAD BH

DCA CH

HLT

AH,

1345

AL,

2167

BH,

0312

BL,

0110

CH,

CL,

A similar procedure is followed to subtract two double precision
numbers. The following program illustrates the technique.

*200

DUBSUB,

CLA CLL

TAD BL

CIA

TAD AL

DCA CL

RAL

DCA KEEP

TAD BH

CMA

TAD AH

TAD KEEP

DCA CH

CLL

HLT

AH,

1345

AL,

2167

BH,

0312

BL,

0110

CH,

CL,

KEEP,

The location KEEP is used to save the contents of the link while the
value of BH was complemented in the accumulator. To form a double
precision two's complement number, a double precision one's comple-

3-15

ment is formed and the 1 is added to it once. Thus, the value of BL is
complemented using the CIA instruction, while the value of BH is
complemented with the CMA instruction. The CLL instruction is used
to clear the link and disregard the carry resulting from using two's com-
plement numbers to perform subtraction.

Powers of Two

In the decimal number system, moving the decimal point right (or
left) multiplies (or divides) a number by powers of ten. In a similar way,
rotating a binary number multiplies (or divides) by powers of two. How-
ever, because of the logical connection between the accumulator and
the link bit, care must be taken that unwanted digits do not reappear in
the accumulator after the passage through the link. Multiplication by
powers of two is performed by rotating the accumulator left; division is
performed by rotating the accumulator right. Multiplication and division
by this method are subject to the limitation of 12-bit numbers (unless
double precision is used). That is, significant bits rotated out of the
accumulator by multiplication or division are lost and incorrect results
are therefore obtained. For example, the following program multiplies a
number by 8 (2 3 ).

*200

MULT8, CLA CLL

TAD NUMBER
CLL RAL
CLL RAL
CLL RAL
DCA NUMBER
HLT
NUMBER, 0231

$
The program will replace the number 023 1 8 by 2310 8 . Notice that
multiplying any number with four significant octal digits (such as
1234 s ) using this program will yield incorrect results.

WRITING SUBROUTINES

Included in the memory reference instructions, given in Chapter 2,
was the instruction JMS (jump to subroutine).. This instruction is a
modified JMP command which makes return to the point of departure
from the main program possible. The JMS instruction automatically
stores the location of the next instruction after the JMS in the location
to which the program is instructed to jump, thereby enabling a return.

3-16

The programmer need only terminate the subroutine with an indirect
JMP to the first location of the subroutine in order to return to the
next instruction following the JMS instruction. The following simple
program illustrates the use of a subroutine to double a number con-
tained in the accumulator.

START,

(Main Program)

CLA CLL
TAD N
JMS DOUBLE
DCA TWON

(Get the number in the AC)
(Jump to subroutine to double N)
(First instruction after the subrou-
tine)

nnnn

TWON,

nnnn

(Subroutine)

DOUBLE,

0000

DCA STORE

TAD STORE

CLL RAL

SNL

JMP I DOUBLE

CLA CLL

TAD STORE

HLT

STORE,

0000

(Any number, N)
(2N will be stored here)

(Save Value of N)
(Get N back in the AC)
(Rotate left, multiplying by 2)
(Did overflow occur?)

(If overflow occurs, display the
number to be doubled in the AC
and then stop the computer.)

Notice that the first instruction of the subroutine is located in the second
location of the subroutine. Any instruction stored in location DOUBLE
would be lost when the return address is stored. Also note that the sub-
routine as it is written must be located on page or current page, be-
cause it is directly addressed. (A subroutine is often located on another
page and addressed indirectly as the next example demonstrates.)

The following program multiplies a number in the accumulator by a
number stored in the location immediately following the JMS instruc-
tion.

3-17

(Main Program)

*200

START,

CLA CLL

TAD A

DCA .+3

TAD B

JMS I 30

0000

DCA PRDUCT

PRDUCT,

0000

0051

0027

*30

MULT

(Subroutine)

*6000

MULT,

0000

CIA

DCA MTALLY

TAD t MULT

ISZ MTALLY

JMP .—2

ISZ MULT

JMP I MULT

MTALLY,

0000

The preceding example illustrates the following important points.

1. The JMS I 30 instruction could be used anywhere in core
memory to jump to this subroutine because the pointer word
(stored in location 30) is located on page and all pages of
memory can reference page 0.

2. The period was used to denote the current location in the in-
structions DCA .-f 3 and JMP .—2.

3. Since the result of the subroutine is left in the AC when jump-
ing back to the main program, the next instruction should store
the result for future use.

4. The first instruction of the subroutine is in location MULT + 1
since the next address in the main program is stored in MULT
by the JMS instruction.

3-18

5. The first two instructions of the subroutine set the tally with
the negative of the number in the AC.

6. The second number to be multiplied is brought into the sub-
routine by the TAD I MULT instruction since it is stored in
the location specified by the address that the JMS instruction
automatically stores in the first location of the subroutine. This
is a common technique for transferring information into a sub-
routine.

7. The ISZ MTALLY instruction is used in the subroutine to
count the number of additions. The ISZ MULT instruction is
used to increment the contents of MULT by one, thereby mak-
ing the return jump (JMP I MULT) proceed to the next in-
struction after the location which held the number to be multi-
plied.

8. An interesting modification of the previous program is achieved
by defining a "new operation" MLTPLY by including in the
coding the statement MLTPLY =JMS I 30. The assembler
would make a replacement such that any time the programmer
writes MLTPLY, the computer would perform a jump to the
subroutine and return to the program with the product in the
AC.

ADDRESS MODIFICATION

A very powerful tool often used by the programmer is address modi-
fication, meaning the inclusion of instructions in a program to modify
the operand portion of a memory reference instruction. It is a particu-
larly useful technique when working with large blocks of stored data
as illustrated by the two programs that follow.

The first program sums 100 8 numbers in locations 300 8 to 377 8 . The
program begins in location 200 8 . The block of 100 8 numbers is sum-
med using only one TAD instruction merely by repeatedly increment-
ing and performing the instruction. ~~

The second example program moves data between memory pages as
well as performing an operation upon the data. The program computes
the square of the 200 8 numbers in locations 4000 8 to 4177 8 . The pro-
gram starts in location 200 8 . All numbers to be squared must not ex-
ceed 45 10 or the square is too large to be represented in the normal
format.

3-19

( START )

[add first number|
-»jaop next number^

ADD I TO THE ADO
INSTRUCTION

Program

START,

CLA CLL

TAD K100

CIA

DCA TALLY

ADD,

TAD 300

ISZ ADD

ISZ TALLY

JMP ADD

DCA SUM

HLT

K100,

0100

TALLY,

0000

SUM,

0000

3-20

The second example illustrates the method of using indirect address-
ing in an address modification situation. It should be noted that in the
first example the actual instruction was incremented to perform the
modification. In the second example, the modification was done by
incrementing the contents of a location which was used for indirect
addressing. The second example could be simplified further through
use of autoindexing, a feature that will be discussed later.

START,

CLA CLL

TAD K200

CIA

DCA TALLY

TAD K4000

DCA NUM

TAD K4200

DCA RESULT

AGAIN,

TAD I NUM

JMS SQUARE

DCA I RESULT

ISZ RESULT

ISZ NUM

ISZ TALLY

JMP AGAIN

HLT

K200,

0200

TALLY,

0000

K4000,

4000

NUM,

0000

K4200,

4200

RESULT,

0000

*300

SQUARE,

0000

DCA STORE

TAD STORE

CIA

DCA COUNT

TAD STORE

ISZ COUNT

JMP -2

JMP I SQUARE

STORE,

0000

COUNT,

0000

3-21

The reader should note that the first eight instructions of the second
example are concerned with intializing the program. This intializing
enables the stored program to be restarted several times and still oper-
ate on the correct locations. If the program had merely incremented
locations K4000 and K4200 and utilized those locations for indirect
addressing the program would only operate on the correct locations of
the first running. On successive runnings the program would be opera-
ting on successively higher locations in memory. With the program
written as shown however the pointer words are automatically reset.
This procedure is often referred to as "housekeeping."

INSERTING COMMENTS AND HEADINGS

Because programs very seldom are written, used, and then forgotten,
the programmer should strive to document his procedure and coding
as much as is reasonably possible. There are many instances where
changes or corrections must be made by people unfamiliar with a pro-
gram, or more commonly the original programmer is asked to modify
a program months after his original effort. In both cases, the success of
the attempt to change the program depends largely upon the documen-
tation provided by the original programmer. A complete and accurate
flowchart is the first form of documentation. It is extremely important
to document modifications made in the program by incorporating these
changes in the flowchart as well.

Many times it is desired to include headings and dates to identify a
program within the actual coding of the symbolic program. It is often
helpful to add comments to simplify the reading of a symbolic program
and to indicate the purpose of any less than obvious instruction. PDP-8
programming allows comments and headings to be inserted simply by
preceding any comments with a slash (/).

The following example illustrates the method used to insert com-
ments and headings in a PDP-8 program. It also illustrates the use of
a rotate instruction. The program takes a binary word stored in memory
and counts the number of non-zero bits. Although the program may
have no useful application, it does serve to familiarize the reader with
the structure of the accumulator and link bit and the action of a rotate
instruction. The flowchart and comments will aid the reader to under-
stand the program.

3-22

( START ~}

TXT

GET THE NUMBER

ROTATE LEFT

NO ^ \ YES

DONE YET ?

( SJ0P - )

/COUNT THE BINARY ONES

/20 SEPTEMBER 1968

*200

START, CLA CLL

DCA COUNT

TAD I WORD

SNA

HLT
ROTATE, RAL

SNL

JMP .—2

CLL

1SZ COUNT

SNA

HLT

JMP ROTATE
COUNT,
WORD, 3000

PROGRAM

/SET COUNT TO 0.
/GET THE WORD.

/STOP IF THE WORD IS 0.
/ROTATE ONE BIT INTO LINK.
/WAS THE BIT = 0?
/YES: ROTATE AGAIN.
/NO: CLEAR LINK.
/COUNT THE NUMBER OF l'S.

/STOP IF THE WORD IS NOW 0.

/ANY 12-BIT NUMBER.

3-23

The following points should be observed in the preceding example.

1. The word was checked to see that it was non-zero to begin
with. If this check were not made, a zero word would be rotated
endlessly by the remaining instructions in the program.

2. Because a rotate right instruction (RAR) would transfer the
bits into the link just as the RAL instruction does, either could
be used in the above program. Both instructions use a circular
shift of the accumulator and link bits.

3. Because the link bit is rotated into the accumulator by the
rotate instructions, the link must be cleared each time a 1 is
rotated into it.

LOOPING A PROGRAM

As many of the examples given have already shown, the use of a
program loop, in which a set of instructions is performed repeatedly,
is common programming practice. Looping a program is one of the
most powerful tools at the programmer's disposal. It enables him to per-
form similar operations many times using the same instructions, thus
saving memory locations because he need not store the same instructions
many times. Looping also makes a program more flexible because it is
relatively easy to change the number of loops required for differring
conditions by resetting a counter. It is good to remember that looping
is little more than a jump to an earlier part of the program; however,
the jump is usually conditioned upon changing program conditions.

There are basically two methods of creating a program loop. The
first method is using an ISZ (2nnn 8 ) instruction to count the number
of passes made through the loop. The ISZ, is usually followed by a JMP
instruction to the beginning of the loop, This technique is very efficient
when the required number of passes through the loop can be readily
determined.

The second technique is to use the Group 2 Operate Microinstruc-
tions to test conditions other than the number of passes which have
been made. Using this second technique, the program is required to
loop until a specific condition is present in the accumulator or link bit,
rather than until a predetermined number of passes are made.

To illustrate the use of an ISZ instruction in a program loop situa-
tion, consider the following program which simply sets the contents of
all addresses from 2000 to 2777 to zero.

3-24

*200

CLEAR,

CLA

TAD CONST

DCA COUNT

/SET COUNT TO -1000.

TAD TTABLE

DCA STABLE

/SET STABLE TO 2000.

DCA I STABLE

/CLEAR ONE LOCATION.

ISZ STABLE

/SELECT NEXT LOCATION.

ISZ COUNT

/IS OPERATION COMPLETE?

JMP .-3

/NO: REPEAT.

HLT

/YES: HALT.

CONST,

7000

/2'S COMP OF 1000.

COUNT,

TTABLE,

TABLE

STABLE,

/POINTER TO TABLE.

*2000

TABLE,

Several points should be carefully noted.

1. The first five instructions initialize the loop, but are not in it.
The location COUNT is set to -1000 at the beginning, and 1
is added to it during each passage of the loop. After the 1000th
(octal) passage, COUNT goes to zero, and the program skips
the JMP instruction, and executes the HLT instruction. On
each previous occasion, it executed the JMP instruction.

2. In the list of constants following the HLT instruction, TTABLE
contains TABLE, which is in turn defined below as having the
value 2000, and containing 0. Therefore, STABLE contains
2000 initially. In order to understand this point it must be re-
membered that an asterisk character causes the first location
after the asterisk to be set to the value after the asterisk. There-
fore, in the previous example CLEAR equals 200 and TABLE
equals 2000.

3. ISZ STABLE adds 1 to the contents of location STABLE,
forming 2001 on the first pass, 2002 on the second pass, and
so on. Since it never reaches zero, it will never skip. This is a
very common use. It is said to be indexing the addresses from
2000 to 2777. (When using an ISZ instruction in this way, the
programmer must be certain that it does not reach 0. Follow
the ISZ instruction with a NOP if necessary. )

3-25

4. For every ISZ instruction used in a program, there must be
two initializing instructions before the loop, and there must be
a constant and a counting location in a table of constants. This
procedure allows the program to be rerun with the counting
locations reset to the correct values.

The following program utilizes a Group 2 skip instruction to create
a loop. The program will search all of core memory to find the first
occurrence of the octal number 1234.

*0
NUMBER,

1234

*200

BEGIN,

CLA CLL
TAD NUMBER
CIA

DCA COMPARE

/STORES MINUS NUMBER.

DCA ENTRY

/SETS ENTRY TO 0.

REPEAT,

ISZ ENTRY
CLA

/INCREASES ENTRY.

TAD I ENTRY

/COMPARISON IS

TAD COMPARE

/DONE HERE.

SZA CLA

JMP REPEAT

TAD ENTRY

HLT

/ENTRY IS IN AC.

COMPARE,

ENTRY,

The previous example is not very useful perhaps but it is interesting
to note that the program will search itself as well as all other core
memory locations.

Also notice the following points with regard to the example.

1. The ISZ ENTRY instruction is used to index the locations to
be tested. The next instruction (CLA) is unnecessary, thus if
ENTRY becomes zero during the course of the program, the
program will not be affected. It is very important to protect
against an ISZ instruction going to zero and skipping a neces-
sary part of a program, if the ISZ is being used to simply
index.

3-26

2. The number to be searched for was stored in location 0, and
the search starts in location 1. Therefore, the program will find
at least one occurrence of the number and will halt after one
complete pass through memory if not before.

3. The program could be modified to bound the area of the search.
By setting the contents of ENTRY equal to one less than the
desired start location and putting the number being searched for
in the location following the last location to be searched, the
program will search only the designated area of memory.

4. The program could be restarted at location REPEAT in order
to find a second occurrence of 1234 after the program had
halted with the first occurrence.

AUTOINDEXING

The PDP-8 family computers have eight special registers in page 0,
namely locations 0010 through 0017. Whenever these locations are
addressed indirectly by a memory reference instruction, the content of
the register is incremented before it is used as the operand of the in-
struction. These locations can therefore be used in place of an ISZ in-
struction in an indexing application. Because of this unique action these
eight locations are called autoindex registers. It is important to realize
that autoindex registers act as any other location when addressed
directly. The autoindexing feature is performed only when the location
is addressed indirectly.

The following example is a modification of the first program example
in the preceding section with an autoindex register used in place of the
ISZ instruction. (The purpose of the program is to clear memory loca-
tions 2000 through 2777. )

Carefully notice the difference between the two examples, especially
that TABLE now has to be set to TABLE-1 since this is incremented
by the autoindexing register before being used for the first time. This

3-27

point must be remembered when using an autoindex register. The
register increments before the operation takes place, therefore it must
always be set to one less than the first value of the addresses to be
indexed.

•10

INDEX,

•200

CLEAR,

CLA

TAD CONST

DCA COUNT

TAD TTABLE

DCA INDEX

DCA I INDEX

ISZ COUNT

JMP -2

HLT

CONST,

7000

COUNT,

TTABLE,

TABLE-1

*2000

TABLE,

The memory search example of the preceding section could also be
simplified using an autoindex register as shown below.

♦0

NUMBER,

1234

•10

ENTRY,

•200

BEGIN,

CLA CLL

TAD NUMBER

CIA

DCA COMPARE

DCA ENTRY

REPEAT,

TAD I ENTRY

TAD COMPARE

SZA

JMP REPEAT

TAD ENTRY

HLT

COMPARE,

Notice that in this case ENTRY
originally equals because its
content is incremented before
being used to obtain data for
the comparison.

3-28

PROGRAM DELAYS

Because the development of a computer was primarily sparked by a
desire for speed in performing calculations, it seems inconsistent and
self-defeating to slow the computer down with program delays. How-
ever, there are many occasions when a computer must be told to slow
down or to wait for further information. This is because most peripheral
equipment, and certainly the human operator, is very much slower than
the computer program. A temporary delay may be introduced into the
execution of a program when needed by causing the computer to enter
one or more futile loops which it must traverse a fixed number of times
before jumping out. It is often necessary to have a computer perform a
temporary delay while a peripheral device is processing data to be sub-
mitted to the computer. The delays can be accurately timed so as not
to waste any more computer time than necessary.

The following is a simple delay routine using the ISZ instruction for
an inner loop and an outer loop. The reader should remember when
analyzing the example that the PDP-8 represents only positive numbers
up to 3777 8 or 2047i O . Therefore, the computer counts up to 2047 10
and then continues to count starting at the next octal number 4000 8 ,
which the computer interprets as — 2048 10 . Successive increments of
this number will finally bring the count to zero. Thus, a location could
be used to count from 1 up to by using an ISZ instruction.

(main program)

TAD CONST

/START OF DI

DCA COUNT

ISZ COUNT1

/INNER

JMP .— 1

/LOOP

ISZ COUNT

jmp -3

CONST,

6030

/SETS DELAY

COUNT,

COUNT1,

The inner loop consists of an ISZ instruction with an execution time
on the PDP-8/I of 3.0 microseconds (a microsecond is 10~ 6 seconds)
and a JMP instruction with an execution time of 1.5 microseconds.
Therefore, the inner loop takes 4.5 microseconds for one pass, and each
time it is entered the program will traverse it 4096 t0 times before leav-
ing. This means that a delay of 18.432 milliseconds (a millisecond is

3-29

10~ 3 seconds) has occurred. If, as in the example above, the value of
CONST is 603 8 , this loop will be entered 1000 10 times giving a total
delay of 18.432 seconds. For any given purpose, a desired delay of
from milliseconds to seconds can be obtained precisely by varying the
values of CONST and the initial value of COUNT 1. Similar reasoning
can be used to design delays for other members of the PDP-8 family.

A second type of delay, which waits for a device response, is dis-
cussed in Chapter 5. This type is not a timed delay but causes the com-
puter to wait until it receives a response from an external device.

PROGRAM BRANCHING

Very few meaningful programs are written which do not take ad-
vantage of the computer's ability to determine the future course the
program should follow based upon intermediate results. The procedure
of testing a condition and providing alternative paths for the program
to travel for each of the different results possible is called branching a
program. The Group 2 microinstructions presented in Chapter 2 are
most often used for this purpose. The ISZ instruction also provides a
branch in a program. These instructions are often referred to as con-
ditional skip instructions. The ISZ instruction operates upon the con-
tents of a memory location, while the Group 2 microinstructions test
the contents of the AC and L.

A typical example of a conditional skip would be a program to com-
pare A and B and to reverse their order if B is larger than A.

SAVE A
IN DUMMY

STORE B IN
A'S LOCATION

STORE DUMMY IN
B'S LOCATION

3-30

*200

TEST,

CLA CLL

TADB

/SUBTRACT B

CIA

/FROM A

TAD A

/HERE.

SMA CLA

HLT

/STOP HERE IF A IS GREATER
OR EQUAL

TAD A

/THE REMAINDER OF

DCA DUMMY

/THE PROGRAM

TADB

/DOES THE SWITCH.

DCA A

TAD DUMMY

DCAB

HLT

1234

/SUBSTITUTE ANY POSITIVE

2460

/VALUES FOR A AND B.

DUMMY,

If A is less than B, their difference will be negative and the HALT will
be skipped. The program will proceed to reverse the order of A and B.
If A is greater than or equal to B, the program will halt.

The concept illustrated by the above example can be included in a
larger program that will take a set of elements and arrange them in
increasing order. The following important concepts should be learned
from the example.

1. The program contains two loops to perform the. sort. The inner
loop starts at TEST and is traversed 20 8 times to switch ad-
jacent elements of the set. The outer loop begins at START
and is re-entered until the elements are in the correct order.

2. A "software flag" was created to signal the program that a
switch has been performed on the last pass. The flag is checked
upon every exit from the inner loop. If the flag is non-zero
(equal to — 1), a reverse was performed on the last pass and
the next pass is started. If the flag is zero, the set is now in
order and the program halts.

3. The flag is set to zero on each pass through the outer loop by
depositing AC=0 in it. It can only be set to- a non-zero value
by a pass through the REVERSE subroutine.

4. The TALLY had to be set to -(AMOUNT) +1 or in this
case to — 20 8 because if the set contains n elements there are
n— 1 comparisons between an element and the immediately
succeeding element, thus, in this case, TALLY=— 20 8 .

3-31

5. The following sort of five elements illustrates the technique
used in the program.

INITIAL

PASS I

PASS 2

©. © © ©

o o
©
©
©
©

In performing the above sort, the program maices three passes.
On the third pass through the table of data, the flag is not
raised; therefore, the program stops.

| SET FLAG ~| — » IN ™ E X M 2 ENT *

3-32

♦200

START,

CLA

CLL

TAD

AMOUNT

CIA

I AC

DCA

TALLY

DCA

FLAG

TAD

BEGIN

DCA

TAD

BEGIN

IAC

DCA

TEST,

TAD
CIA

I X2

TAD

I XI

SPA

SNA CLA

SKP

JMS

REVERSE

ISZ

TALLY

JMP

TEST

TAD

FLAG

SZA

JMP

START

HLT

AMOUNT,

TALLY,

0000

BEGIN,

2000

XI,

0000

X2,

0000

FLAG,

0000

HOLD,

0000

REVERSE

,0000

TAD

I XI

DCA

HOLD

TAD

I X2

■ DCA

I XI

TAD

HOLD

DCA

I X2

CLA

CLL CMA

DCA

FLAG

JMP

I REVERSE

/THESE INSTRUCTIONS SET

/UP A TALLY EQUAL TO

/AMOUNT -1 TO COUNT THE

/PASSES THRU TEST LOOP.

/CLEARS FLAG BEFORE EACH PASS

/THESE INSTRUCTIONS

/SET THE POINTERS

/Xl AND X2 TO THE

/PROPER VALUES

/INITIALLY.

/SUBTRACTION FOR THE

/TEST IS

/DONE HERE.

/DO SWITCH IF AC IS POSITIVE.
/SET UP THE X'S FOR
/THE NEXT PASS-
/HAVE ALL X'S BEEN TESTED?
/NO! KEEP TESTING. •
/YES: WERE ANY SWITCHES
/DONE ON THE LAST PASS?
/YES: GO THRU PROGRAM AGAIN.
/NO: STOP, TABLE IS' IN ORDER.

/SUBROUTINE TO SWITCH X'S

/SETS AC EQUAL TO - 1 .
/SET FLAG=-1 ON A SWITCH.

6. This program can perform a sorting for any specified block of
data merely by specifying the octal number of entries to be
sorted in the location AMOUNT and by specifying the begin-
ning address of the block in BEGIN. The data to be sorted
must be placed in consecutive memory locations.

3-33

Exercises,

1. Write a subroutine SUB to subtract the number in the AC from
the number in the location after the JMS instruction that calls the
subroutine. Return to the main program with the difference in the
AC. Use a flowchart and comments to document the procedure.

2. Write two programs to put into memory location 2000, 1 into

2001, 2 into 2002, etc., up to 777 8 into 2777 using (a) an ISZ
instruction for indexing and (b) autoindexing. Use flowcharts
and comments to document the procedure.

3. The following program was previously given to multiply two num-
bers together.

*200

START, CLA CLL

TAD A

CIA

DCA TALLY
MULT, TAD B

ISZ TALLY

JMP MULT

HLT
A 1

g ' J- substitute any numbers for A and B

TALLY, 0000

a. What is the largest product that the PDP-8 can compute using
this program?

Using the following value for A and B, verify that the program will
obtain the correct answers. Remember that any carry from the most
significant bit is lost from the accumulator.

Ax B

b.
c.
d.

7756(-18 1 „)

0000
7700 (— 64 10 )

0027 (23 10 )
0005
0000

4. Write a program TRIADD which will add two triple precision
numbers A+B=C. There are three parts to each number, namely
AH (A high) AM (A medium), and AL (A low); BH, BM, and
BL; CH, CM, and CL. Use a flowchart and comments to docu-
ment the procedure.

3-34

5. Write a program to perform a multiplication between two single-
precision numbers to yield a double-precision product. Use com-
ments and a flowchart to document the procedure.

6. Write a program to multiply any number n by a power of 2 (the
exponent is stored in location EXP), the product being expressed
in double precision. Use comments and a flowchart to document
the procedure.

7. Write a program to find how many of the numbers stored in a
table from address 3000 to address 3777 are negative. Use a flow-
chart and comments to document the procedure.

8. Write a program that will run for exactly 20 seconds on the PDP-8
or PDP-8/I before it halts. Use a flowchart and comments to
document the procedure.

9. Modify the program written for exercise 8 such that if bit 11 of
the console switch register is a 1, the program runs for 20 seconds,
and if it is a 0, the program runs for 40 seconds.

Hint: The OSR instruction must be used to check the switch
register.

10. The program on the next page rotates a bit left or right depend-
ing on the value of bit and faster or slower depending on the
value of the remaining bits. Analyze the program and comment
each instruction to indicate its use in the program.

3-35

*200

ROTATE,

CLA CLL CML

HLT

BEGIN,

DCA SAVEAG

RAL

DCA SAVEL

TAD MASK

OSR

DCA COUNT

OSR

RAL -

SZL CLA

JMS LEFT

JMS RIGHT

CLL

GO,

TAD SAVEL

RAR

TAD SAVEAC

INSTR,

RAR

ISZ COUNTR

JMP .-1

ISZ COUNT

JMP -3

JMP BEGIN

SAVEAC,

SAVEL,

MASK,

7000

COUNTR,

COUNT,

LEFT,

ISZ LEFT

TAD KRAL

DCA INSTR

JMP I LEFT

RIGHT,

TAD KRAR

DCA INSTR

JMP I RIGHT

KRAR,

7010

KRAL,

7004

3-36

Chapter 4

System Description
and Operation

The PDP-8 system is composed of the computer console, a Teletype
console (usually an Automatic Send Receive Model 33) and possibly
other peripheral equipment. While normal operation of a computer sys-
tem is by programmed control, manual operation is necessary for many
tasks. This chapter describes the manual control and operation of the
PDP-8/I and PDP-8/Lspecifically. as representative of the PDP-8 com-
puter family. This chapter also provides an introduction to the more
common peripheral devices which may be included in a PDP-8 system.
Chapter 5 describes the programmed control of peripheral devices and
the means for transferring information between peripheral equipment
and the central processor.

CONSOLE OPERATION

The operator console allows manual control of the computer and
provides the most elementary means of storing a program in memory.
It is a collection of switches and indicator lamps which enable the pro-
grammer to examine the contents of locations in memory, alter the
contents of memory locations, and determine the current status of a
running program.

Console Components

The PDP-8/I operator console is shown in Figure 4-1. The PDP-8/L
operator console is significantly different from the PDP-8/I console
and is shown in Figure 4-2. The 'following discussion applies to both
console diagrams except where differences are noted. For reference
purposes, the switches and indicators are identified in the following
tables.

4-1

Switch
POWER

PANEL LOCK

START

LOAD ADD

DEP

EXAM

CONT

Explanation

This key-operated, switch controls the com-
puter's primary power supply. This switch has
"on" and "off" positions. In the PDP-8/L,
however, the switch includes a third position
which performs the function of the PANEL
LOCK switch, below.

This key-operated switch disables all console
switches except the switch register. (This fea-
ture is provided in the PDP-8/L by the third
position of the POWER switch.)
The START switch initiates execution of the
computer program beginning in the location
currently held in the PROGRAM COUNTER
register. The START switch clears the ACCU-
MULATOR, LINK, MEMORY BUFFER
and instruction registers.
The LOAD ADD (load address) switch sets
the contents of the switch register (SR) into
the PROGRAM COUNTER (PC). (The PC
is an internal register which is not displayed
on the PDP-8/L console.) The computer
should be halted before operating this switch.
The DEP (deposit) switch sets the contents of
the switch register into the location currently
specified in the PROGRAM COUNTER. De-
pressing this switch also results in the PRO-
GRAM COUNTER being incremented by 1,
and the address of the location previously
specified by the PC is displayed in the MEM-
ORY ADDRESS register. Successive opera-
tion of the switch stores contents of the switch
register in successive locations because the
PROGRAM COUNTER is incremented each
time the DEP switch is depressed.
The EXAM (examine) switch displays the
contents of the location specified by the
PROGRAM COUNTER in the MEMORY
BUFFER and ACCUMULATOR displays.
Depressing this switch also results in the PRO-
GRAM COUNTER being incremented by 1;
thus, repeated manipulation of the EXAM
switch displays the contents of successive
memory locations.

The CONT (continue) switch initiates execu-
tion of the stored program at the location
specified by the PROGRAM COUNTER. The
CONT switch does not clear any active regis-
ters. (The START switch does.)

4-2

i :a.v»»»«*» «BM i > » V ^' ' » *-»• ai« » vir * <T»i ,^ M**- . . •»t** „ > r v~ •m . ft-

rrr*- ' ■" ' t «-

Figure 4-1. PDP-8/I Computer Console

Figure 4-2. PDP-8/L Computer Console

4-3

Switch

Explanation

STOP
SING STEP

SING INST

SWITCH REGISTER
(SR)

MEM PROT

DATA FIELD (DF)

and
INST FIELD (IF)

The STOP switch halts program execution at
the end of the instruction in progress.
When the SING STEP (single step) switch is
set, the computer executes instructions one
memory cycle at a time for each depression of
the CONT switch.

When the SING INST (single instruction)
switch is set, the computer will execute one
instruction at a time for each depression of the
CONT switch. This switch is not present on
the PDP-8/PDP-8/L console.
The SWITCH REGISTER is a set of twelve
toggle switches used to specify binary numbers
which are loaded into registers when other
console switches are operated. LOAD ADD
sets the contents of the SR into the PRO-
GRAM COUNTER. DEP sets the contents of
the SR into memory through the MEMORY
BUFFER register. The twelve positions repre-
sent a 12-bit binary word.
PDP-8/ 1 When the top of a switch is out, it
represents a binary 1 and is con-
sidered set; conversely, when the
bottom of the switch is out it rep-
resents a binary and is not set.
PDP-8/L, When the switch is up, it repre-
PDP-8/S, sents a binary 1 and is considered
PDP-8 set; conversely, whm the switch is
down it represents a binary and
is not set.
This switch is provided on the PDP-8/ L only.
When set (depressed), the MEM PROT
(memory protect) switch prevents the storing
or changing of information in the upper 200 8
locations (7600 to 7777) of core memory (the
upper 200„ locations of field 1 in an 8K com-
puter) .

NOTE: JMS, DCA, ISZ instructions and
all input transfers are not per-
mitted, thus protecting programs
or data stored in this area.
These two console elements (switches and dis-
plays) are enabled and used when the basic
computer is equipped with extended memory.
Their function is described later in this chap-
ter.

4-4

Indicator Display

PROGRAM
COUNTER
(PC)

MEMORY ADDRESS

(MA)

MEMORY BUFFER

(MB)

ACCUMULATOR
(AC)

LINK (L)

MULTIPLIER
QUOTIENT

(MQ)
Instruction and
Status Indicators

Explanation

The contents of this 12-bit display represent
the address of the next instruction to be exe-
cuted. This display is not present on the PDP-
8/ L console.

Contents represent the address of the word
being obtained from or stored in memory.
After depressing the DEP or EXAM switches,
the contents represent the address of the word
previously read or written.
The contents of this 12-bit display represent
the word currently being obtained from or
stored in memory.

This display indicates the current contents of
the ACCUMULATOR. After depressing the
DEP or EXAM switch, this display indicates
the contents of the location whose address is
displayed in the MA.

This display indicates the contents of the
LINK, a 1-bit register which serves as an ex-
tension of the ACCUMULATOR.
This display is activated by the EAE option
described later in this chapter. The option is
not available on the PDP-8/L.
These indicators are located in the upper right
of the console. The display indicates the status
of the program being executed and the opera-
tion code of the instruction being executed.

Manual Program Loading

Having written a program, the programmer must store the instruc-
tions in memory before they can be executed. The octal value of the
instructions may be directly loaded into memory from the computer
console. Once the POWER switch has been turned on to energize the

4-5

computer, the programmer loads the instructions into memory using
the console switches. The location in which the instruction is to be
stored is identified by setting the SWITCH REGISTER to the desired
address, and then operating the LOAD ADD switch. The address will
be placed in the program counter register and displayed in the PRO-
GRAM COUNTER indicator.

To load an instruction or data into the location specified by the
PROGRAM COUNTER, a similar procedure is followed. The
SWITCH REGISTER is set to the binary representation of the instruc-
tion (or data) and then the DEP switch is depressed. The instruction
(or data) is displayed in the MB and AC displays and the address of
the location is displayed in the MA display. Loading a program into
sequential memory locations is simplified by the fact that the DEP
switch will automatically increment the program counter register by 1.
Thus, once an initial address is specified, instructions and data may be
loaded into sequential memory locations by alternately setting the binary
representations on the SWITCH REGISTER and operating the DEP
switch.

Once loaded into memory, a program may be checked by using the
EXAM switch. To check the contents of a location, the desired address
is entered on the SWITCH REGISTER and the LOAD ADD switch
is depressed. The EXAM switch is used to display the contents of the
specified location in the MEMORY BUFFER display. Because opera-
tion of the EXAM switch will increment the program counter by 1 (as
^loes the DEP switch), sequential locations may be examined by re-
peated use of the EXAM switch, once the initial location is specified
using the SWITCH REGISTER and LOAD ADD switch. Thus, a
stored program may be checked to see that it was correctly loaded by
displaying each program location.

To run a stored program, the binary value of the starting address of
the program is toggled on the SWITCH REGISTER and then the
LOAD ADD is depressed. (The starting address appears in the PC.)
Depressing the START switch causes the computer to execute the pro-
gram beginning with the instruction specified by the address in the
PROGRAM COUNTER register. A program may be manually halted
by the operation of the STOP switch.

Console switch positioning and procedures for initializing the console
are outlined below. Figures 4-3 through 4-5 summarize the procedures
for loading, checking, and running a program.

4-6

INITIALIZE

SET SR TO

PROGRAM'S FIRST

ADDRESS

DEPRESS LOAD ADO
SWITCH

SET SR TO

PROGRAM'S FIRST

INSTRUCTION

DEPRESS DEP
SWITCH

SET SR TO

PROGRAM'S NEXT

INSTRUCTION

YES

PROGRAM

IS LOADED

CONSOLE SWITCH POSITIONING
PDP-8/I When the top of a switch is out,
it represents a binary 1- and is con-
sidered set; conversely, when the
bottom of the switch is out it rep-
resents a binary and is not set.
PDP-8/L, When the switch is up it represents
PDP-8/S, a binary 1 and is considered set;
PDP-8 conversely, when the switch is
down it represents a binary and
is not set.

INITIALIZING THE CONSOLE

1. Computer POWER is on.

2. PANEL LOCK is off.

3. SWITCH REGISTER equals zero.

4. Both SING STEP and SING INST are
not set.

5. All peripheral devices turned off.

Figure 4-3. Manually Loading a Program

4-7

INITIALIZE
4

•

SET SR TO

THE FIRST
ADDRESS

DEPRESS

LOAD ADD

SWITCH

DEPRESS EXAM
SWITCH

SET SR TO
VALUE OF MA

CONTENTS OF MA

IS DISPLAYED

IN MB

DEPRESS

LOAD ADD

SWITCH

' CORRECT \,

,. INSTRUCTION ^S

SET SR
TO CORRECT
INSTRUCTION

|YES

DEPRESS DEP
SWITCH

.•""ALL^N.
^INSTRUCTIONS^N.

PROGRAM
IS CHECKED

Figure 4-4. Checking a Stored Program

( INITIALIZE

)

SET SR TO

ADDRESS

OF FIRST

INSTRUCTION

DEPRESS LOAD ADD
SWITCH

DEPRESS START
SWITCH

(PROGRAM "\
IS RUNNING J

Figure 4-5. Running a Stored Program

4-8

TELETYPE OPERATION

The ASR 33 Teletype console is the basic input/output device for
the PDP-8 computer family. It consists of a printer, keyboard, paper
tape reader, and paper tape punch. The Teletype unit can operate
under program control or under manual control. Programmed operation
of the Teletype unit is described in detail in Chapter 5. Operation of the
Teletype unit as an independent device for generating paper tapes is
described later in this section.

Teletype Unit Components

The ASR 33 Teletype unit commonly used in conjunction with the
PDP-8 computer family is pictured in Figure 4-6 with the major features
noted.

OFF
LINE O L0CAL

Figure 4-6. ASR 33 Teletype Console
4-9

The components of the Teletype unit and their functions are de-
scribed in the following paragraphs.

CONTROL KNOB

The control knob of the ASR 33 Teletype console (see Figure 4-6)
has the following three positions.

LINE The Teletype console is energized and connected to the

computer as an input/ output device under computer con-
trol.

OFF The Teletype console is de-energized.

LOCAL The Teletype console is energized for off-line operation
under control of the Teletype keyboard and switches ex-
clusively.

O ©©©©'© O © © © © O ©

§)©©©©©©©©©©©©

©O©®©©OO©©©©0©

0©O©O©©®©©©0

Figure 4-7. Teletype Keyboard

KEYBOARD

The Teletype keyboard shown in Figure 4-7 is similar to a type-
writer keyboard, except that some nonprinting characters are included
as upper case elements. For typing characters or symbols, such as $,
% , #, which appear on the upper portion of numeric keys and certain
alphabetic keys, the SHIFT key is held depressed while the desired
key is operated.

Designations for certain (nonprinting) operational functions are
shown on the upper part of some alphabetic keys. By holding the
CTRL (control) key depressed and then depressing the desired key,
these functions are activated. Table 4-1 lists several commonly used
keys that have special functions in the symbolic language of PDP-8
family computers.

4-10

PRINTER

The printer provides a typed copy of input and output at ten char-
acters per second maximum rate. When the Teletype unit is online
(LINE), the copy is generated by the computer; when the Teletype
unit is offline (LOCAL), the copy is automatically generated when-
ever a key is struck.

PAPER TAPE READER

The paper tape reader is used to input into memory data punched
on eight-channel perforated paper tape at a maximum rate of ten
characters per second. The reader control positions are shown in Fig-
ure 4-6 and are described below.

START Activates the reader; reader sprocket wheel is engaged and
operative.

STOP Deactivates the reader; reader sprocket wheel is engaged

but not operative.
FREE Deactivates the reader; reader sprocket wheel is disengaged.

Table 4-1. Special Keyboard Functions

Key

Function

Use

SPACE
RETURN
HERE IS
RUBOUT
CTRL/REPT/P

LINE FEED

space used to combine and delimit symbols

or numbers in a symbolic program

carriage return used to terminate line of symbolic
program

blank tape used for leader/ trailer (effective only

in LOCAL)

rubout used for deleting characters, punches

all channels on paper tape

code 200 used for leader/ trailer of binary pro-

gram paper tapes (keys must be re-
leased in reverse order: P, REPT,
CTRL)

line feed follows carriage return to advance

printer one line

4-11

PAPER TAPE PUNCH

The paper tape punch is used to perforate eight-channel rolled oiled
paper tape at a maximum rate of ten characters per second. The punch
controls are shown in Figure 4-6 and described below.

REL. Disengages the tape to allow tape removal or loading.

B. SP. Backspaces the tape one space for each firm depression of

the B. SP. button.
ON Activates the paper tape punch.

OFF Deactivates the paper tape punch.

CHANNELS

7 6 5 4

3 2 1

SPROCKET
HOLE

Data is recorded (punched) on paper
tape by groups of holes arranged in a
definite format along the length of the
tape. The tape is divided into channels,
which run the length of the tape, and
into columns, which extend across the
width of the tape, as shown in the ad-
jacent diagram. The paper tape read-
ers and punches used with PDP-8
family computers accept eight-channel
paper tape.

Generating a Symbolic Tape

The previously described components may be used to generate a
symbolic program paper tape through the following procedure.

When switched to LOCAL, the Teletype unit is independent of the
computer and functions like an electric typewriter. Any character
struck on the keyboard is printed, and also punched on paper tape if
the tape punch is ON. Each character struck on the keyboard is rep-
resented in code by one row of holes and spaces according to the
ASCII code described in the following section and given in Appendix B.

4-12

A section of leader-trailer code several inches long is punched at
the beginning of the symbolic tape, by pressing the HERE IS key on
the Teletype keyboard. The symbolic program is then carefully typed,
following the conventions used in PDP-8 symbolic programs as de-
scribed in Chapter 3.

A typing error can be corrected using the B.SP. button of the paper
tape punch and the RUBOUT key on the Teletype keyboard. The B.SP.
button backspaces the paper tape one column for each depression of
the button, and the RUBOUT key perforates all eight channels of a
column (this perforation is ignored by the computer). Therefore,
errors are removed by backspacing the tape to the error and typing
rubouts over the error and all following characters. After typing rub-
outs, the correct infomation must be typed beginning where the error
occurred.

Once the symbolic tape is punched, more leader-trailer tape is gen-
erated by striking the HERE IS key. The tape is removed from the
punch unit by tearing against the plastic cover of the punch. The sym-
bolic program thus generated is the input to the assembler described
in Chapter 6.

The program may be listed (typed out) by placing the paper tape
in the paper tape reader. This is done by releasing the plastic cover
of the reader unit and placing the eight-channel tape over the reader
head with the smaller sprocket holes over the sprocket wheel, and
replacing the cover. If the Teletype control is switched to LOCAL and
the reader is switched to START, the tape will advance over the reader
head and a printed copy of the program will be typed on the Teletype
printer. If the tape punch is also ON, a duplicate of the tape will be
generated at the same time.

Paper Tape Formats

Manual use of the toggle switches on the operator console is a
tedious and inefficient means of loading a program. This procedure is
necessary in some instances, however, because the PDP-8 family of
computers must be programmed before any form of input to the
memory unit is possible. For example, before any paper tape can be
used to input information into the computer, the memory unit must
have a stored program which will interpret the paper tape format for
the computer. This loader program must be stored in memory with
the console switches. A loader program consists of input instructions
to accept information from the Teletype paper tape reader and instruc-
tions to store the incoming data in the proper memory locations.

■ 4-13

Before the loader program can be written to accept information,
the format in which the data is represented on the paper tape must
be established. There are three basic paper tape formats commonly
used in conjunction with PDP-8 family of computers. The following
paragraphs describe and illustrate these formats.

324

3 10
31 1
323
240
31 I
323
240
301
323
303
31 1
31 1
240
306
317
322
315
301
324

ASCII FORMAT

The USA Standard Code for Informa-
tion Interchange (ASCII) format uses
all eight channels 1 of the paper tape to
represent a single character (letter,
number, or symbol) as shown in the
diagram at left. The complete code is
given in Appendix B.

CHANNEL 7

• • • •

77
61
32
12
77
52
I 3
1 1
77
53
13
10
77
00
77
07
07
00

RIM (READ IN MODE)
location FORMAT

contents RIM format tape uses adjacent col-
umns to represent 1 2-bit binary infor-

location ma tion directly. Channels 1 through 6
are used to represent either address or
information to be stored. A channel 7
punch indicates that the adjacent col-
umn and the following column are to

contents be interpreted as an address specifying
the location in which the information

location f t ne following two columns is to be
stored. The tape leader and trailer for

contents jy^j format tape must be punched in

location channel 8 only (octal 200).

location

1 Channel 8 is normally designated for parity check. The Teletype units used
with PDP-8 family computers do not generate parity, and Channel 8 is always
punched.

4-14

• • • •

00
17

ORIGIN

52
50

INSTRUCTION

00
00

INSTRUCTION

00
00

INSTRUCTION

02
56

INSTRUCTION

77
01

INSTRUCTION

76
00

INSTRUCTION

77
60

• INSTRUCTION

01
77

• INSTRUCTION

INSTRUCTION

BIN (BINARY) FORMAT

BIN format tape is similar to RIM
format except that only the first ad-
dress of consecutive locations is speci-
fied. An address is designated by a
channel 7 punch and information fol-
lowing an address is stored in sequen-
tial locations after the designated
address, until another location is speci-
fied as an origin. The tape leader/
trailer for BIN format tape must be
punched in channel 8 (octal 200)
only.

Paper Tape Loader Programs

The three previously described paper tape formats are each used
for a separate purpose in conjunction with PDP-8 family computers.
The ASCII format is used to represent symbolic programs on paper
tapes, which are then used as input to the assembler. As described in
Chapters 2 and 3, the assembler translates the mnemonic instructions
and symbolic addresses into binary instructions and absolute addresses.
Once this translation has been performed by the assembler, a binary
format tape is generated.

The binary format tape is the common means of loading an as-
sembled program into the core memory of a PDP-8 family computer.
The BIN (Binary) loader is the program used to load these binary
format paper tapes. Program instructions are stored in successive lo-
cations beginning with an origin which is signaled by a channel 7 punch
on the paper tape. The BIN loader is a lengthy program requiring 83
memory locations. As an alternative to manually entering the contents
of all 83 locations, the RIM (Read In Mode) format is used.

The RIM loader is simpler than the BIN loader because the memory
unit is supplied with a location for each incoming instruction. It con-
sists of 17 instructions which must be toggled into memory. The BIN
loader is punched in RIM format, and is loaded by the RIM loader;
but it is used to load tapes punched in the BIN format, which is the
output of the assembly program.

The RIM loader is listed in Table 4-2. The instructions are toggled
in, and checked by following the flowcharts given in Figures 4-3 and

4-15

4-4. The instructions are given for use with both the low speed reader
(included in the Automatic Send Receive Model 33 Teletype) and
for the high speed reader (an option described later in this chapter).

Table 4-2. RIM (Read In Mode) Loader

Instruction

Location

Low-Speed Reader

High-Speed Reader

7756

6032

6014

7757

6031

6011

7760

5357

7761

6036

6016

7762

7106

7763

7006

7764

7510

7765

5357

5374

7766

7006

7Q06

7767

6031

6011

7770

5367

7771

6034

6016

7772

7420

7773

3776

7774

3376

7775

5356

5357

Note: Location 7776 is used for temporary storage.

The RIM loader will load into memory any program punched on
paper tape in the RIM format. Because paper tapes in the RIM and
BIN formats cannot be used until the user understands the material
in Chapter 6, further discussion of the use of paper tape loaders is
contained in that chapter.

PERIPHERAL EQUIPMENT AND OPTIONS

PDP-8 family computers are used in many different environments
and are interfaced with many different peripheral devices. The Teletype
unit is the most common peripheral device, but other equipment and
options often incorporated in a system with the PDP-8 include high
speed paper tape reader and punch units, DECtape, DECdisk, ex-
tended memory, and the extended arithmetic element (EAE). These
options give the basic PDP-8 new capabilities of which the program-
mer should be aware. The purpose and features of each of these op-
tions is described in the following paragraphs.

All of the options listed above and many others may be directly
connected to the PDP-8/ 1. The PDP-8/ L may be directly equipped
with the high speed paper tape reader and punch unit only. Other

4-16

peripherals must be interfaced to the PDP-8/L through an I/O con-
version panel and/ or a peripheral expansion panel. EAE is not avail-
able with the PDP-8/L.

High Speed Paper Tape Reader and Punch Unit

Loading a long paper tape program into the PDP-8 core memory
with the low-speed reader of the ASR 33 Teletype unit is very time
consuming. Punching a long program on paper tape from an assembly
program likewise is very slow, If handling lengthy paper tapes is com-
monly required, much computer time is wasted while these low-speed
I/O devices read or punch data. The high-speed paper tape reader
and punch unit, shown in Figure 4-8, performs paper tape input and
output at a considerably faster rate. It is of great value in a system
that relies on paper tape as a primary medium of data and program
storage.

Figure 4-8. High-Speed Paper Tape Unit

The high-speed paper tape reader is used to input data into core
memory from eight-channel, fan-folded (non-oiled), perforated paper
tape. The reader inputs information photoelectrically at a rate of 300
characters per second (ASR 33 reader inputs at ten characters per
second maximum). Primary power is applied to the reader when the
computer console POWER switch is on. The reader is controlled by
the computer, although the operator may indirectly control the reader
from the keyboard through the computer. Tape may be advanced,
without being recorded by the photoelectric sensors, by pressing the
white advance button.

4-17

Paper tapes are manually positioned in the high speed reader with
the following steps.

1. The paper tape is placed in the right-hand bin such that the
beginning of the tape will pass over the sensors first.

2. Several folds of leader tape are placed in the left-hand bin
with the tape passing under the tape retainer cover.

3. The retainer cover is closed over the tape such that the feed
holes are engaged in the teeth of the sprocket wheel.

4. Tape is advanced and read by programmed computer in-
structions.

Once the paper tape has been properly placed in the reader and the
leader/ trailer has been positioned as outlined in the preceding steps,
the tape is normally read under control of system software.

The high speed paper tape punch is used to record computer out-
put on eight-channel, fan-folded paper tape at 50 characters per sec-
ond. All characters are punched under program control from the com-
puter. Primary power is available to the punch when the computer
console POWER switch is turned on. Power is supplied when the
POWER button is depressed on the punch unit itself. In addition to
the POWER button, a FEED button is located on the punch enclosure
to advance feed-hole-only punched tape for leader/ trailer purposes.

The loader programs, symbolic assemblers, Symbolic Editor, and
other system software presented in Chapter 6 include instructions for
using the high speed reader/ punch, as well as for using the Teletype
reader/punch. (By incorporating the appropriate instructions for the
high speed unit (see Appendix D), the user may write his own I/O
routines for this device as outlined in Chapter 5. )

Extended Memory

The PDP-8 family of computers have a memory unit composed of
12-bit magnetic core locations. The basic configuration stores 4,096
12-bit words; however, the memory unit of the PDP-8/ 1, -8/S, or -8
can be expanded into a maximum storage of 32,768 words by adding
4,096-word memory modules. The PDP-8/ L may be expanded to
8,192 words. Each module is called a field, with field being the
original 4,096 words and other fields designated 1, 2, . . ., 7. (The
PDP-8/L can have only fields and 1.)

Expansion of the basic memory introduces data field and instruction
field registers into the memory unit. Related to these registers are the
DATA FIELD and INST FIELD switches and displays of the com-
puter console. Since the 12-bit word of the PDP-8 is capable of rep-
resenting only 4,096 locations uniquely, the data field and instruction

4-18

field registers are used to designate the field of 4,096 words which
contains a particular address.

INSTRUCTION FIELD

The content of the instruction, field register determines the instruc-
tion field (field of 4,096 words) that the instructions are to be taken
from. Any directly addressed AND, TAD, ISZ, or DCA instruction
will obtain its operand from the instruction field. In indirectly ad-
dressed instructions, however, the pointer address is taken from the
instruction field, but the operand (specified by the effective address)
is obtained from the data field.

DATA FIELD

The content of the data field register specifies the data field (field of
4,096 words) from which operands (specified by the effective address)
are taken in indirectly addressed AND, TAD ISZ or DCA instructions.
(The pointer addresses are obtained from the instruction field.)
INITIAL FIELD ASSIGNMENTS

The original setting of the data field and instruction field registers
are by the DATA FIELD and INST FIELD switches of the computer
console (see Figure 4-1). Thus, to run a program beginning in loca-
tion 200 of field 1 and operating on data in field 0, the INST FIELD
(IF) switches are set to 001, and the DATA FIELD (DF) switches
set to 000 (on the PDP-8/L, set to 1 and respectively). The switch
register is then set to 200 s . When the LOAD ADD switch is operated,
the values for the data field and instruction field are entered as well
as the starting address. When the START switch is depressed, the pro-
gram beginning in location 200 of field 1 is executed.

A common use of extended memory is the storage of system soft-
ware. For example, the Binary ^Loader may be stored in field 1. By
setting IF = 1 and DF = 0, the Binary Loader runs in field 1, but de-
posits the program (which is simply data to the loader) in field 0.

CHANGING FIELD ASSIGNMENTS

The instructions for extended memory (see Appendix D) may be
used to change instruction and data fields during the execution of a
program. The instructions are written in the following format.

Symbolic Instruction Explanation

CIF+30 change to instruction field 3

CDF+ 1 change to data field 1

The field being changed to is specified by adding its value to the second
octal digit position of the change field instruction. The CDF instruction

4-19

causes all future indirectly-addressed operands to be taken from the
specified data field. The CIF (change instruction field) instruction does
not take effect immediately, but waits for a JMP or JMS instruction to
be encountered in the program execution. When the JMP or JMS is
encountered, control is transferred to the new instruction field. Thus,
if the instruction field is originally field 0, the following instructions
transfer control to field 1 .

CIF+10
JMP 20

The next instruction to be executed in the program is contained in
location 20 of field 1. If the JMP is indirectly-addressed, the pointer
word is obtained from the old field, but the JMP is to a location in the
new field.

DECtape System

DECtape is an option of the PDP-8 family of computers which
serves as an auxiliary magnetic tape storage facility and updating de-
vice. The standard DECtape transport unit is pictured in Figure 4-9.
The DECtape system stores and retrieves information at fixed positions
on magnetic tape. The advantage of DECtape over conventional mag-
netic tape is that information is stored at fixed positions which may be
addressed. Allocation of fixed, addressable positions for information
storage is a unique feature of DECtape storage facilities, while con-
ventional magnetic tape stores information in sequential (not directly
addressable), variable-length positions. DECtape incorporates timing
and mark information to reference the fixed positions. The 10-channel
DECtape records five channels of information: a timing channel, a
mark channel, and three information channels. These five channels are
duplicated on the remaining five nonadjacent channels to minimize
any possibility of loss of information from the other channels. The
DECtape is organized in blocks of data words with control words to
identitfy each block. The tape is bidirectional; that is, it can be written
or read in the forward or reverse direction. Information should be read
in the same direction that it was written.

The DECtape control unit performs the transfer of information be-
tween the PDP-8 and the transport unit. The control can operate as
many as eight separate DECtape transport units.

4-20

Figure 4-9. DECtape Transport Unit

The DECtape transport unit shown in Figure 4-9 is a bidirectional
magnetic tape transport which reads and writes the 10-channel mag-
netic tape. Tape movement can be controlled by programmed instruc-
tions from the computer or by the manual operation of switches located
on the front panel of the transport. Data is transferred only under pro-
gram control.

The transport controls are identified below.

Transport Control
REMOTE

OFF

LOCAL

WRITE ENABLED
WRITE LOCK

Unit Selector

Explanation

This switch position energizes the DECtape
transport and places it under program control.
This switch position disables the DECtape
transport.

This switch position energizes the DECtape
transport and places it under operator control
from external transport switches.
This switch position enables the DECtape for
search, read, and write activities.
This switch position limits the DECtape trans-
port to search and read activities only. (This
prevents accidental destruction of permanent
data.)

The value specified by this eight-position ro-
tary switch identifies the transport to the con-
trol unit.

NOTE: Position 8 on the Unit Selector

switch corresponds to DECtape

unit 0.

4-21

With the transport in LOCAL mode, depress-
ing this switch causes tape to feed onto the
right-hand spool.

With the transport in LOCAL mode, depress-
ing this switch causes tape to feed onto the
left-hand spool. The REMOTE and WRITE
ENABLED lamps indicate whenever their re-
spective conditions are present.

DECdisk System

The DECdisk file is a fast, random-access, bulk storage device for
the PDP-8 computer family. It has a considerably faster access time
than DECtape and offers storage of 32,768 words on each disk (as
many as four disks are possible). The DECdisk comprises a storage
unit with electronics to perform the read and write functions, and the
computer interface logic which participates in the transfer of informa-
tion between the central PDP-8 and the DECdisk unit.

The storage unit contains a motor-driven, nickel/cobalt-plated disk.
The disk has 16 data tracks, with 2,048 words per track. There are
two timing tracks plus two spares. Data is recorded on a single disk by
fixed position read/write heads. Transfer of information between the
DECdisk and the PDP-8 is controlled by programmed instructions as
outlined in Chapter 5.

A fast, convenient, keyboard-oriented Disk/DECtape Monitor is
available for use with the PDP-8 family computers to allow the pro-
grammer to efficiently control the flow of programs through any PDP-8
having a DECdisk or DECtape. This monitor system is described in
more detail in Chapter 7.

Extended Arithmetic Element

The EAE option of the PDP-8 computer family (not available on
the PDP-8/L) provides circuitry to perform arithmetic operations which
can not be directly performed with the basic PDP-8 instruction set.
The option includes microinstructions to perform multiplication and
division. Other microinstructions perform arithmetic and logical shifts
and normalize both positive and negative two's complement numbers.
The option provides a 12-bit multiplier quotient register (MQ)
which is used in conjunction with the AC to perform direct multiplica-
tion and division. The content of this register is displayed on the PDP-8
console.

The EAE option is essentially an increase in instruction capability.
The instructions, which are microprogrammable, are included in Ap-
pendix D.

4-22

NOTE TO READER

The following exercises are intended for readers who have access to
a computer of the PDP-8 family. Readers who do not presently have
access to such a computer should begin study of Chapter 5.

Upon completion of the following exercises, readers with access to
a PDP-8 would benefit from reading the sections of Chapter 6 which
describe loaders, the Symbolic Editor, and the PAL III Symbolic
Assembler. Although knowledge of this material is not necessary to
understand the subject matter of Chapter 5, this knowledge will facili-
tate the running of programs presented therein.

EXERCISES

1. Toggle into memory and run the programs written in Chapter 2,
for exercises 6 and 10.

2. Toggle into memory the RIM Loader (Table 4-2) using the con-
sole switches. Verify the contents of the registers with the EXAM
switch.

3. Write a program to set the contents of locations 2000 through
2007 to the value of the switch register and then halt. Toggle it in
and verify that it works.

4. Write a program to accept two numbers from the switch register
and add them displaying their contents in the accumulator. (Hint:
precede each OSR instruction by a HLT. After seeing the switch
register activate the CONT key.) Translate the program into octal
and toggle it into memory. Verify that it works properly.

4-23

4-24

Chapters

Input /Output
Programming

Being able to program a computer to do calculations is of little use
if there is no way of getting the results of calculations from the
machine. Likewise, the programmer often must supply the computer
with information to be processed. A programmer must be provided
with the means to transfer information between the computer and the
peripheral devices that supply input or that serve as a means of output.

Before a transfer of information can be executed, a control function
must be supplied to specify when the exchange will occur, with what
peripheral device the exchange will occur, and where in core storage
the information will be stored (or obtained from). In general, this con-
trol function may be served by either the PDP-8 or the peripheral de-
vice itself.

There are three basic methods for the transfer of information be-
tween input/output (I/O) devices and the PDP-8. The first two
methods provide for PDP-8 control over the transfer. One method is
programmed transfer, in which instructions are included at some point
in the program to accept or transmit information. Thus, programmed
transfers are program initiated and are under program control.

Information may also be transferred through program interrupt, a
standard feature of the PDP-8 computer family that provides for de-
vices to signal the PDP-8 when they are ready to transfer information;
the program will then interrupt its normal flow and jump to a routine
to process the information, after which it will return to the point in the
main program at which it was interrupted. Thus, program interrupt
transfers are device initiated but are under program control.

5-1

These first two methods (i.e., programmed transfers and program
interrupt) use the accumulator as the buffer, or storage area in the
computer, for all data transfers. Therefore, only one 12-bit word of
input or output may be transferred at one time by a programmed
transfer, or by program interrupt.

The third method of information transfer is data break, an option
for the PDP-8/L but standard for other PDP-8 computers. Data break
is essentially device controlled and allows for direct exchange of large
quantities of information between the device and the PDP-8 memory.
It differs from the previous two types of transfer in that there are no
program instructions to handle the transfer and the accumulator is not
used as a buffer. Data break transfers are device initiated and device
controlled.

INPUT/OUTPUT INSTRUCTIONS

As the name implies, programmed transfers of information are ac-
complished with a set of program instructions. The instructions are
similar to the operate microinstructions in that there is no need to
specify an address in memory. The operation code 6 8 is used to specify
an input/ output transfer (IOT) instruction. All programmed transfers
are between the accumulator and the device. Since many different de-
vices could be connected to one computer and each device may at
some time transfer information, the instruction must identify the proper
device for each transfer. The instruction must also specify the exact
nature of the function to be performed.

IOT Instruction Format

An IOT instruction is a 12-bit word that is in the following format.
The first three bits represent the operation code 6 8 . The remaining nine
bits may be either binary O's or l's.

/ \

OPERATION
CODE

DEVICE SELECTION
CODE

OPERATION

SPECIFICATION

BITS

Figure 5-1. The IOT Instruction
5-2

The IOT instruction is divided into three parts: operation code, de-
vice selection code, and operation. specification bits.

Device Selection

The device selection code is transmitted to all peripheral equipment
whenever the IOT instruction is executed. A device selector within
each peripheral device monitors the device codes. When the device
selector recognizes a device code as the device's assigned code, the de-
vice receives the last three bits of the instruction. Each of the last three
bits specifies an action associated with the device. When one of the last
three bits is set to a 1, the specified action is performed. Since there
are three bits, only three different actions can be specified for each de-
vice code, although microprogramming is possible. When more instruc-
tions are necessary for a given device, more than one code is assigned
to the device.

Checking Ready Status

Because there is a great difference in the processing speed of a com-
puter and the speed of most peripheral devices, the computer must
check the readiness of a device before any transfer of information is
performed. The input device must signal the computer that it has com-
pletely assembled the information and is now ready to transfer the in-
formation to the computer memory. The output device must signal its
readiness to accept the next piece of information from the computer.
Without such signals, the computer would input and output informa-
tion at a faster rate than the device could process it and some infor-
mation would be lost.

To prevent any loss of information, the computer program checks
the ready status of the transmitting or receiving device as part of pre-
paring for a normal data transfer. The ready status is usually checked
with a skip instruction such that if the device is ready, the following
instruction is skipped. The ready status is signaled through a system
of flags, which are 1-bit registers within the device . All I/O devices
have a device flag which is set to a 1 when the device is ready; that is,
when it can be used (if it is an output device), or when it has informa-
tion (if it is an input device). If the flag is cleared (set to 0), the device
is busy. If a program initiates a device action, the flag associated with
that device will be set to a 1 when the device action is completed.

5-3

Instruction Uses

In general, for each device there are three instructions:

1. An instruction to transfer information and operate the device.

2. An instruction to test the ready status of the device and skip
on the ready (or not-ready) status of the device.

3. An instruction to clear the device flag.

The above instructions may be microprogrammed. In particular, the in-
structions to clear the flag and to operate the device often are com-
bined.

The specific instructions for devices are given in the following
sections. The Teletype unit is described in depth to explain the fun-
damentals of programming data transfers. The general techniques
developed for the Teletype unit may be extended to handle other de-
vices.

ASCII Code

The ASCII (U.S.A. Standard Code for Information Interchange) is
presented in Table 5-1. Many of the programs written in this chapter
use this code to transmit information to the PDP-8. The fact that* the
ASCII code for the octal digits through 7 is the sum of that digit
plus 260 a should be observed.

PROGRAMMING THE TELETYPE UNIT

One of the most common I/O devices is the Teletype unit, which
contains a keyboard, printer, paper tape reader, and paper tape punch.
The Teletype unit can use either the keyboard or the paper tape reader
to input information to the computer and can use either the printer or
the paper tape punch to accept output information from the computer.
The Teletype unit is therefore assigned two device codes.

Teletype Inpnt/Ontpnt Transfer Instructions

Functioning as an input device, the keyboard/reader is assigned the
device code 03 8 , and functioning as an output device, the printer/
punch is assigned the device code 04 8 .

5-4

Table 5-1. The 8-BM ASCII 1 Code

8-Bit

Character

Octal

Character

Octal

301

. 241

302

242

303

243

304

$ .

244

305

245

306

246

307

247

310

(

250

311

)

251

- J

312

252

313

253

314

254

315

255

316

256

317

/ ■

257

320

272

321

273

322

274

323

—

275

324

276

325

277

326

300

327

[

333

330

334

331

]

335

332

336

260

337

261

Leader/Trailer

200

262

LINE FEED

212

263

Carriage RETURN

215

264

SPACE

240

265

RUBOUT

377

266

Blank

000

267

BELL

207

270

TAB

211

271

FORM

214

1 An abbreviation for USA Standard Code for Information Interchange.

5-5

KEYBOARD/READER INSTRUCTIONS

The instruction format for the keyboard/reader is shown in Figure
5-2. The mnemonic instructions generated by bits 9, 10, and 11 are
noted. The sequence in which the mnemonic instructions are executed
when microprogrammed is noted below.

OPERATION
CODE

Figure 5-2. Teletype Keyboard/Reader Instructions

Sequence Mnemonic Octal

2,3

KSF

KCC
KRS

KRB

6032
6034

6036

Effect

Skip the next instruction when

the keyboard buffer register is

loaded with an ASCII symbol

(causing the keyboard flag to be

raised).

Clear AC, clear keyboard flag.

Transfer the contents of the key-
board buffer into the AC.
Transfer the contents of the key-
board buffer into the AC, clear
the keyboard flag.

The fourth instruction (KRB) is a microprogrammed combination of
the mnemonics KCC and KRS. If the paper tape reader is loaded with
a paper tape and switched to START, the KRB instruction accepts one
character from the reader.

A program using the above instructions to read in one ASCII char-
acter from the keyboard or paper tape reader is shown in Figure 5-3.
Note that this program does not type the character on the teleprinter, it
merely stores the ASCII code for the character in the location STORE.

5-6

*200
INPUT,

LISN,

STORE,

KCC

JMS LISN
DCA STORE
HLT

KSF

JMP .-1

KRB

JMP I LISN

/CLEAR KEYBOARD FLAG

/SKIP ON KEYBOARD FLAG
/READ KEYBOARD BUFFER

Figure 5-3. Coding to Accept One ASCII Character

The main program begins with KCC. In general, the main program
should begin by clearing the flags of all devices to be used later in the
program. If the above program is started at location 200, it will pro-
ceed to the KSF, JMP .—1 loop, and stay in this loop endlessly until a
key on the Teletype unit is pressed or a paper tape is loaded into the
reader. When the ASCII code for the character is assembled in the
keyboard/reader buffer register, the flag will be set to a 1 and the pro-
gram will skip out of the loop. The contents of the buffer will be trans-
ferred into the accumulator, and the buffer and flag will be cleared.

PRINTER/PUNCH INSTRUCTIONS

The instruction format for the Teletype printer/punch IOT instruc-
tions is given in Figure 5-4. The mnemonic instructions generated by
bits 9, 10, and 1 1 are discussed on the following page.

.OP

-^ A

ERATION

DEVICE CODE
4

TSF
TCF

TPC

Figure 5-4. Teletype Printer/Punch Instructions
5-7

Sequence
1

Mnemonic
TSF

Octal
6041

2
3

TCF
TPC

6042
6044

2,3

TLS

6046

Effect
Skip the next instruction if the
printer flag is set to 1 .
Clear the printer flag.
Load the printer buffer register
with the contents of the AC, se-
lect and print the character. (The
Hag is raised when the action is
completed.)

Clear the printer flag, transfer
the contents of the AC into the
printer buffer register, select and
print the character. (The flag is
raised when the action is com-
pleted.)

The last instruction is a microprogrammed combination of TPC and
TCF, such that the flag is cleared, the character is printed, and then
the flag is again raised. Whenever the paper tape punch is turned on,
the character is punched on paper tape as well as printed on the
teleprinter.

Figure 5-5 illustrates a program to print out one ASCII character
which is held in a memory location.

*200
OUTPUT,

TYPE,

HOLD,

CLA CLL

TLS

TAD HOLD

JMS TYPE

HLT

TSF

JMP.-l

TLS

JMP I TYPE

/SKIP ON TELEPRINTER FLAG
/PRINT THE CHARACTER

Figure 5-5. Coding to Print One ASCII Character

The program in Figure 5-5 begins by clearing the accumulator and
executing a TLS instruction (which has the effect of clearing the printer
buffer), after which the printer flag will be set, thereby signifying
readiness to accept a character. If the initial TLS instruction were not

5-8

executed, the flag would not be raised (the START key clears all flags),
and the program would remain in the TSF, JMP .—1 loop endlessly.
In the previous case, however, the program uses the printer with a cleared
accumulator such that no character is printed. However, the flag is set
when this action is complete enabling the printing of meaningful infor-
mation in the TYPE subroutine. The TYPE subroutine clears the
accumulator since the TLS instruction does not. It is advisable to clear
the accumulator after any subroutine unless meaningful data is con-
tained in it.

Format Routines

Input and output routines are very often written in the form of sub-
routines, as the TYPE subroutine in the previous example. The exam-
ple in Figure 5-6 is a carriage return/line feed subroutine that calls
the TYPE subroutine to execute a carriage return and line feed on the
printer, thus advancing to a new line for the printing of information.

CRLF,

TAD K215

JMS TYPE

TAD K212

JMS TYPE

JMP I CRLF

K215,

215

K212,

212

TYPE,

TSF

JMP .-1

TLS

CLA CLL

JMP I TYPE

/ASCII FOR CARRIAGE RETURN
/ASCII CODE FOR A LINE FEED

Figure 5-6. Carriage Return/Line Feed Subroutine

Subroutines similar to the one in Figure 5-6 could be written to tab
space the carriage a given number of spaces, or to ring the bell of Tele-
type Model ASR 33 by using the respective codes for these nonprinting
characters. Such subroutines, if commonly used in a program, should
be placed on page (or else a pointer word to the subroutine should
be placed on page zero) to facilitate reaching the routine from all
memory locations.

5-9

Text Routines

The examples in Figures 5-3 and 5-5 may be expanded to accept
and type more than one character. Figures 5-7 and 5-8 illustrate one
expansion. These two programs are compatible in that the characters
accepted by the first program are typed out by running the second pro-
gram. The program to accept characters, in Figure 5-7, will continue
to accept character input until a dollar sign ($) is struck on the key-
board, at which time the program will store all zeros in the next loca-
tion and then halt. The program in Figure 5-8 will type the characters
whose ASCII code was stored by the first program. The second pro-
gram will halt when a location with contents equal to zero is reached.
Both programs use locations beginning with 2000 as the buffer for the
storage of ASCII characters. The following flowcharts introduce the
techniques used in the program coding.

START

SET BUFFER POINTER
TO FIRST LOCATION

CHECK FLAG
AGAIN

STORE ONE ,

CHARACTER IN BUFFER

PRINT THE
CHARACTER

INCREASE THE
BUFFER POINTER

EXECUTE TLS

TO CLEAR BUFFER

AND SET FLAG

SET BUFFER POINTER
TO FIRST LOCATION

RETURN
CARRIAGE

GET NEXT
ASCII CODE

/ IS IT ^N^
< ZERO ]>

Two

TYPE OUT
CHARACTER

INCREASE BUFFER
POINTER

5-10

*5000

START,

CLA

CLL

TAD

BUFF

DCA

BUFFPT

LISN,

KSF

JMP

.-1

KRB

TLS

DCA

I BUFFPT

TAD

I BUFFPT

TAD

MDOLAR

SNA

JMP

DONE

ISZ

BUFFPT

JMP

LISN

DONE,

CLA

CLL

DCA

I BUFFPT

HLT

BUFF,

2000

BUFFPT,

MDOLAR,

7534

/SET UP BUFFER SPACE. '

/KEYBOARD STRUCK YET?
/NO: CHECK AGAIN.
/YES: READ CHARACTER.
/ACKNOWLEDGE IT ON PRINTER.
/STORE CHARACTER.
/CHECK FOR TERMINAL $.

/CHARACTER IS A £.
/CHARACTER IS NOT A S.
/GET ANOTHER CHARACTER.
/STORE IN LAST LOCATION.

Figure 5-7. Routine to Accept and Store ASCII Characters

*5200

START,

CLA
TLS

CLL

TAD

BUFF

DCA

BUFFPT

JMS

CRLF

CHRTYP,

TAD
SNA
HLT

I BUFFPT

JMS

TYPE

ISZ

BUFFPT

JMP

CHRTYP

CRLF,

TAD

K21 5

JMS

TYPE

TAD

K212

JMS

TYPE

JMP

I CRLF

TYPE,

TSF

JMP

• -1

TLS

CLA

JMP

I TYPE

BUFF,

2000

BUFFPT,

K21 5,

21 5

K212,,

212

/TLS TO SET PRINTER FLAG.
/SET UP BUFFER SPACE.

/RETURN CARRIAGE.

/GET A CHARACTER.

/IS IT ALL ZEROS?

/YES: STOP.

/NO: TYPE OUT THE CHARACTER.

/INCREMENT BUFFER POINTER.

/TYPE ANOTHER CHARACTER.

/CARRIAGE RETURN & LINE FEED.

/TYPE CR, FIRST.

/TYPE LINE FEED.

/SUBROUTINE TO TYPE CHARACTER.
/PRINTER READY YET?
/NO: CHECK AGAIN.
/YES: TYPE CHARACTER.
/CLEAR ASCII FROM AC

Figure 5-8. Routine to Print Stored ASCII Characters
5-11

The program to print characters may be specialized to print a
specific word as in the program of Figure 5-9. The example is a
subroutine which uses autoindcx registers in place of the ISZ instruc-
tion. The subroutine types "HELLO!"

HELLO.

NEXT,

TYPE,

CHARAC,

M6,

COUNT,

IRl=10

CLA
, TLS
TAD
DCA
TAD
DCA
TAD
JMS
ISZ
JMP
JMP

TSF
JMP
TLS
CLA
JMP

3!0

30 5
314

31 4
31 7
241
-6

CLL

CHARAC

IR1

COUNT

I IR1

TYPE

COUNT

I HELLO

I TYPE

/HELLO SUBROUTINE

/TLS TO SET PRINTER FLAG.

/SET UP INDEX REGISTER

/FOR GETTING CHARACTERS.

/SET UP COUNTER FOR

/TYPING CHARACTERS.

/GET A CHARACTER.

/TYPE IT.

/DONE YET?

/NO: TYPE ANOTHER.

/YES: RETURN TO MAIN PROGRAM.

/TYPE SUBROUTINE

/USED AS INITIAL VALUE OF IR1

/£

Figure 5-9. Routine to Print One Word

Numeric Translation Routines

The ASCII codes of octal numbers may be transmitted to the PDP-8
memory as in the program of Figure 5-3. However, the ASCII code
for the number must be converted to true octal representation before
the computer may use this input. For example, 6 is represented by the
ASCII code 266. When the Teletype key is struck for 6, the code 266
is transmitted to the computer upon execution of the KRB instruction.
To remove the 260 from the coded number to obtain the octal number
itself, two methods could be used.

5-12

The first method is to examine the binary form of the ASCII code.

000 010 110 110

Setting the first eight bits to zero by using the AND instruction with
the appropriate mask results in the binary value for 6. The appropriate,
mask for this purpose is 17 8 as shown below.

Instruction Operation Comment

000 010 110 110

AND MASK 000 000 001 111
000 000 000 110

ASCII Code 266 in

accumulator

MASK: 17 8

Contents of accumulator after

AND instruction is executed.

The second method of stripping an ASCII coded number is to sub-
tract 260 8 from the character code. The instruction TAD M260 is used
for this operation as shown in the following example.

Instruction Operation

TAD M260

000 010 110 110
111 101 010 000
000 000 000 110

Comment

ASCII Code 266 in accumulator
M260; 7520, (2's comp of 260)
Contents of accumulator after
TAD instruction is executed.

Programs to accept and store an octal digit, using the LISN sub-
routine previously given, are shown in Figure 5-10.

/LISN SUBROUTINE

LISN,

KSF
JMP -1
KRB
JMP I LISN

Figure 5-10. Two Methods of Converting ASCII to Binary

When an octal digit stored in memory is to be printed on the tele-
printer, the octal number 260 must be added to it. The routine in Fig-
ure 5-1 1 uses the TYPE subroutine previously introduced to print out
the binary number 7.

*200

/USING AND

/USING TAD

/INSTRUCTION

NUMIN,

KCC

NUMIN, KCC

JMS LISN

AND MASK

TAD M260

DCA HOLD

HLT

HOLD,

MASK,

M260, 7520

5-13

*200

/TYPE SUBROUTINE

NUMOUT,

CLA CLL

TYPE,

TLS

TSF

TAD NUMBER

JMP .— 1

TAD K260

TLS

JMS TYPE

CLA CLL

HLT

JMP I TYPE

NUMBER,

K260,

260

Figure 5-11. Routine to Type One Stored Digit

The routines presented thus far have been designed to handle only
single-digit octal numbers. The PDP-8 core storage location, how-
ever, is able to represent octal numbers up to four digits. The routine
in Figure 5-12 takes a number that is stored in a PDP-8 location and
prints it on the printer as four octal digits. The routine in Figure 5-13
accepts four octal digits from the keyboard and converts them to one
octal number and stores it in a location, after which it accepts another
four digits, etc. The following flowcharts illustrate the procedure used
in the programs.

STORE DIGiT
"TTMPORaRSLY

BHiNG STORED RESULT
INTO AC

BRING FIRST DIGiT
(NTQ AC

ROTATE LEFT
THREE PLACES

SUBTRACT

ADO NEXT

260 a

DIGiT

STORE THE RESULT
TEMPORARILY

■i BUFFE

5-14

*200

START,

CLA

CLL

DCA

STORE

TLS

JMS

CRLF

TAD

DCA

DIGCTR

TAD

NUMBER

RAL

UNPACK,

TAD
RAL
RTL

STORE

DCA

STORE

TAD

STORE

AND

MASK7

TAD

K260

JMS,

TYPE

ISZ

DIGCTR

JMP

UNPACK

JMS

CRLF

HLT

TYPE,

TSF"

JMP

.-1

TLS

CLA

JMP

I TYPE

CRLF,

TAD

K21 5

JMS

TYPE

TAD

K212

JMS

TYPE

JMP

I CRLF

NUMBER*

. 1234

MASK7,

M4,

-A

DIGCTR,

■

STORE,

K212,

212

K21 5,

21 5

K260,

2 60

/CLEAR OUT STORAGE LOCATION.

/TLS TO SET PRINTER FLAG.

/RETURN CARRIAGE.

/SET LOCATION TO COUNT NUMBER

/OF TYPED DIGITS.

/GET NUMBER TO BE TYPED.

/ROTATE ONE PLACE (INTO LINK).

/ADD STORED LOCATION TO AC-

/ROTATE THREE

/PLACES LEFT.

/STORE ROTATED NUMBER.

/MASK OUT ALL BUT

/FIRST THREE BITS OF NUMBER,

/ADD IN 260,

/AND TYPE DIGIT.

/TYPED FOUR DIGITS YET?

/NO: GO TYPE ANOTHER

/YES: RETURN CARRIAGE

/AND HALT.

/TYPE SUBROUTINE

/CARRIAGE RETURN & LINE FEED

/NUMBER TO BE TYPED

/ASCII FOR LF
/ASCII FOR CR

Figure 5-1 2. Routine to Type a 4-Digit Number
5-1^

*200

START,

CLA

TLS

CLL

TAD

Kl 777

DCA

IR1

JMS

CRLF

NXTNUM,

TAD

DCA

COUNT-R

TAD

K350

DCA

TEMP

NXTDIG,

JMS

LISN

DCA

I TEMP

ISZ

TEMP

ISZ

COUNTR

JMP

NXTDIG

JMS

CRLF

JMS

PACK

DCA

I IR1

JMP

NXTNUM

PACK,

DCA

STORE

TAD

DCA

COUNTR

TAD

K350

DCA

TEMP

PAKDIG,

TAD

STORE

CLL

RAL

RTL

TAD

I TEMP

TAD

M260

DCA

STORE

ISZ

TEMP

ISZ

COUNTR

JMP

PAKDIG

TAD

STORE

JMP

I PACK

CRLF,

TAD

K215

JMS

TYPE

TAD

K212

JMS

TYPE

JMP

I CRLF

/TLS TO SET PRINTER FLAG.
/SET INDEX REGISTER FOR
/STORING PACKED NUMBERS.
/RETURN CARRIAGE.
/SET COUNTER FOR A DIGITS.

/SET UP TEMPORARY STORAGE

/FOR THE ASCII INPUTS.

/GET CHARACTER FROM KEYBOARD.

/STORE IT TEMPORARILY.

/INCREMENT STORAGE LOCATION.

/GOT A DIGITS YET?

/NO: GET ANOTHER.

/YESs RETURN CARRIAGE

/AND PACK THE A DIGITS.

/STORE PACKED NUMBER.

/GET A NEW NUMBER.

/PACK SUBROUTINE.

/CLEAR OUT STORAGE LOC

/SET COUNTER FOR

/A DIGITS.

/SET POINTER TO

/ASCII INPUT CHARACTERS.

/GET STORE IN AC

/ROTATE INTO CLEARED LINK.

/ROTATE IT TWICE MORE.

/ADD NEXT STORED CHARACTER.

/STRIP OFF THE 260.

/STORE STRIPPED NUMBER.

/INCREMENT POINTER TO ASCII.

/PACKED A DIGITS?

/NO: PACK NEXT DIGIT.

/YES: TAKE PACKED NUMBER

/BACK TO MAIN PROGRAM.

/CARRIAGE RETURN LINE FEED.

Figure 5-13. Routine to Pack and Store 4-Digit Octal Numbers

5-16

LISN,

TYPE,

K1777,

IR1=10

M4,

COUNTRj

K350,

TEMP,

STORE,

M260,

K215,

K212,

KSF

JMP .-1

KRB

TLS

JMP I LISN

TSF

JMP .-1

TLS

CLA

JMP I TYPE

1 777

7774

3 50

7520

215

212

/SUBROUTINE TO ACCEPT ASCII

/SUBROUTINE TO TYPE ASCII.

/TEMPORARY STORAGE C350-353>

Figure 5-13. (cont.) Routine to Pack and Store 4-Digit Octal Numbers

Suapfe Program

The previously described routines for typing text and numeric trans-
lation are combined in the following program example which is similar
to the final program of Chapter 3. This program performs the same
numeric sort; however, the numbers to be placed in order are supplied
from the keyboard.

Any number of elements may be supplied; the end of input is sig-
naled by typing a dollar sign ($). The program includes routines to
exclude any nonoctal digits from input and type a question mark. Only
positive octal numbers (0-3777 8 ) are allowed as input to the program.

The program is presented in four illustrations. The following flow-
chart diagrams the program.

5-17

ACCEPT ASCII
CODE FOR
ONE DIGIT

TYPE "?"
AND IGNORE
THAT ENTRY

PUT THE

NUMBERS IN

INCREASING ORDER

PACK AND
STORE AS
ONE NUMBER

— O

Figure 5-14A. Sample Program: Flowchart
5-18

*200

/INITIALIZATION
START, CLA CLL

TLS

TAD BUFF

DCA BUFFPT

DCA AMOUNT
/ACCEPT ONE DIGIT.
ACCEPT, JMS CRLF

TAD M4

DCA DIGCTR

TAD TEMPI

DCA TEMP
NEWDIG, JMS LISN

DCA I TEMP
/CHECK THE CHARACTER.
CHECK, TAD I TEMP

TAD MDOLAR

SNA CLA

JMP ORDER

TAD I TEMP

TAD M260

SPA

JMP ERROR

TAD M10

SMA CLA

JMP ERROR

ISZ TEMP

ISZ DIGCTR

JMP NEWDIG
/YES: PACK THE A DIGITS
PACK, TAD TEMPI

DCA TEMP

DCA HOLD

TAD MA

DCA DIGCTR
DIGPCK, TAD HOLD

CLL RAL

RTL

TAD I TEMP

TAD M260

DCA HOLD

ISZ TEMP

ISZ DIGCTR

JMP DIGPCK
,TAD HOLD

DCA I BUFFPT

TAD I BUFFPT

TAD K4000

SMA CLA

JMP ERROR

ISZ AMOUNT

ISZ BUFFPT

JMP ACCEPT

/TLS TO SET PRINTER FLAG.
/SET UP STORAGE AREA.

/SET AMOUNT TO 0.

/RETURN CARRIAGE.
/SET UP COUNTER
/FOR A DIGITS.
/SET *A POINTER TO
/TEMPORARY INPUT STORAGE.
/GET A CHARACTER.
/STORE IT.

/IS CHARACTER AS?

/YES: ORDER INPUT.

/NO: CHECK FOR OCTAL INPUT.

/IS ASCII LESS THAN 260?

/YES: ERROR.
/NO: SUBTRACT .10.

/ASCII IS GREATER THAN 267.

/INCREMENT STORAGE POINTER.

/A DIGITS YET?

/NO: GET ANOTHER.

INTO ONE NUMBER.

/SET POINTER TO STORAGE LOC

/CLEAR LOCATION HOLD.
/SET COUNTER FOR A DIGITS.

/CONTENTS OF HOLD INTO AC

/ROTATE INTO CLEARED LINK.

/ROTATE TWICE MORE.

/ADD ONE ASCII CHARACTER.

/SUBTRACT OUT THE 260.

/STORE AC IN HOLD.

/INCREMENT STORAGE POINTER.

/PACKED A DIGITS YET?

/N0: PACK ANOTHER.

/YES: STORE PACKED NUMBER.

/NEGATIVE INPUT?

/YES: REJECT ENTRY.
/NO: COUNT THE ENTRIES.
/SET UP FOR A NEW ENTRY.
/GET A NEW ENTRY.

Figure 5-14B. Sample Program: Initialization and Input Coding

5-19

/PUT THE NUMBERS IN INCREASING ORDER.

ORDER, TAD

AMOUNT

CIA

IAC

DCA

TALLY

DCA

FLAG

TAD

BUFF

DCA

TAD

BUFF

IAC

DCA

TEST, TAD

I X2

CIA

TAD

I XI

SMA

SZA CLA

JMS

REVERSE

ISZ

TALLY

JMP

TEST

TAD

FLAG

SZA

CLA

JMP

ORDER

JMP

REVERSE,

TAD

I XI

DCA

HOLD

TAD

I X2

DCA

I XI

TAD

HOLD

DCA

I X2

CLA

CLL CMA

DCA

FLAG

JMP

I REVERSE

/SET UP A TALLY

/TO COUNT THE

/NUMBER OF

/COMPARISONS.

/CLEAR THE FLAG.

/SET THE POINTERS

/CXI AND X2) TO THE

/PROPER DATA LOCATIONS.

/X2=X1+1

/COMPARE XI AND X2.

/REVERSE ENTRIES IF
/X2 IS LESS THAN XI .
/INCREMENT THE POINTERS.

/DONE COMPARING YET?

/NO: COMPARE MORE ENTRIES.

/YES: IS FLAG SET?

/YES: MAKE ANOTHER PASS.
/NO: TYPE THE ORDERED DATA.
/SUBROUTINE TO SWITCH X'S.

/SET FLAG WHENEVER
/A SWITCH IS MADE.

Figure 5-I4C. Sample Program: Ordering Coding
5-20

/PRINT OUT THE ORDERED

PRINT, JMS CRLF
TAD BUFF
DCA BUFFPT
TAD AMOUNT
CIA
DCA PRNTCT

ANOTHR, JMS CRLF
TAD M4
DCA DI6CTR
DCA HOLD
TAD I BUFFPT
CLL RAL

MORE, TAD HOLD
RAL
RTL

DCA HOLD
TAD HOLD
AND MASK7
TAD K260
JMS TYPE
ISZ DIGCTR
JMP MORE
ISZ BUFFPT
ISZ PRNTCT
JMP ANOTHR
JMS CRLF
JMP START

ERROR, CLA

TAD QUEST
JMS TYPE
JMP ACCEPT

END=.

NUMBERS
/RETURN THE CARRIAGE-
/SET THE BUFFER POINTER.

/SET LIMIT FOR OUTPUT.

/RETURN CARRIAGE.

/COUNT THE DIGITS OUTPUT.

/CLEAR HOLD LOCATION.

/GEf A CHARACTER.

/ROTATE INTO CLEARED LINK.

/ADD HOLD TO AC

/ROTATE THREE TIMES LEFT.

/STORE AC IN HOLD-

/MASK OUT FIRST 9 BITS.

/TYPE OUT ONE DIGIT.

/TYPED A DIGITS?

/NO! TYPE ANOTHER-

/YES! INCREMENT BUFFER LOC

/TYPED ALL ENTRIES?

/NO: TYPE ANOTHER ENTRY.

/YES! RETURN CARRIAGE AND

/ACCEPT MORE NUMBERS TO SORT.

/ DISREGARDS ILLEGAL ENTRY.

Figure 5-1 4D. Sample Program: Output Coding
5-21

*100

TYPE;

TSF

JMP

.-1

TLS

CLA

JMP

I TYPE

CRLF,

TAD

K215

JMS

TYPE

TAD

K212

JMS

TYPE

JMP

I CRLF

LISN,

KSF

JMP

.-1

KRB

TLS

JMP

I LISN

BUFF;

END

BUFFPT,

M4,

777^

DIGCTR,

TEMPI,

.+2

TEMP,

MDOLAR,

753^

M10,

-10

K4000,

4000

HOLD,

M260,

-260

AMOUNT,

FLAG,

TALLY,

XI ,

X2,

PRNTCT,

MASK7,

K260,

260

K212,

212

K21 5,

21 5

QUEST,

277

/TYPE OUTPUT SUBROUTINE.

/CARRIAGE RETURNfiLINE FEED.

/LISN INPUT SUBROUTINE.

Figure 5-14E. Sample Program: Subroutines and Constants Coding

5-22

PROGRAM INTERRUPT FACILITY

The running time of programs using input and output routines is
primarily made up of the time spent in waiting for the device to accept
or transmit information. Specifically, this time is spent in loops such as:

TSF
JMP.-l

Waiting loops waste a large amount of computer time. In those cases
where the computer can be doing something else while waiting, these
loops can be removed and useful routines can be included to use this
waiting time. This sharing of a computer between two tasks is often
accomplished through the program interrupt facility, which is standard
on all PDP-8 family computers.

The following two instructions control the interrupt facility.

Mnemonic Octal Operation

ION 6001 Turn interrupt facility on.

IOF 6002 Turn interrupt facility off.

The program interrupt facility allows a running program to proceed
until a peripheral device connected to the interrupt facility sets its
ready flag. The running program is often referred to as the background
program. Whenever a flag is set to 1 by a device that is connected to
the interrupt facility, the PDP-8 completes execution of the instruction
in progress and then acknowledges the interrupt. The interrupted com-
puter will automatically execute a JMS instruction. The result of this
action is that the program counter register, which contains the address
of the next instruction to be performed in the main program, is stored
in location 0. The instruction in location 1 is then performed, which
usually initiates a service routine for the peripheral device.

The service routine, sometimes called the foreground program, is
usually contained elsewhere in memory and an indirect • jump to the
start of the program is contained in location 1. The service routine is
terminated by a JMP I instruction, to return to the background pro-
gram.

The occurrence of an interrupt disables the facility from further in-
terrupts. The ION instruction must be included in the interrupt service
routine to enable further interrupts. This' is usually done immediately
before returning to the background program. The ION instruction does
not take effect until one instruction following it has been completed.

5-23 ■

Thus, the ION instruction is usually followed by the JMP I 0.
INTRTE,

SERVICE
ROUTINE

EXIT,

ION

Enables interrupt facility

JMP I

again after execution of

JMP I 0.

0000,

0202

0001,

JMP INTRTE

0200,

ISZ COUNT

0201,

TAD A

Interrupt request oc-
curred during the execu-
tion of this instruction;
caused a JMS to be ex-
ecuted immediately after
completion of the TAD
A instruction execution.

0202,

RAL

Interrupt is turned on
before this instruction is
executed.

Programming an Interrupt

The program presented in Figure 5-15 includes a program to rotate
one bit through the accumulator as the background program. The fore-
ground program, which is initiated by the service routine, accepts
ASCII characters from the Teletype unit and, upon receipt of the
ASCII code for a period, prints out the characters which have been
stored.

The program begins with an initialization routine to set up buffer
space to store the incoming characters and to set the mode for input.
(The program signals input mode by a value of MODE = and out-
put mode by a value of MODE =1.) Once the initialization routine
has enabled the interrupt facility, the background program is started.

The background program is a routine to rotate one bit through the
accumulator and link. The first instruction clears all bits but bit 1 1. The
program then counts through the two ISZ loops, after which it rotates
the bit one place left and then returns to the count loops. The accumu-

5-24

lator and link displays will exhibit a quickly rotating light while waiting
for an interrupt to initiate the foreground program.

The first instruction to be executed after an interrupt request is an
automatic JMS 0, thereby storing the return address to the background-
program in location 0. The program then executes the instruction in
location 1 which is an indirect jump to the service routine location 2
which contains SERV).

Since an interrupt occurs during the running of the background pro-
gram, the service routine must save the active registers of the back-
ground program. The service routine stores the link and accumulator,
so that they may be restored after the service routine is completed.

The service routine must determine the source of the interrupt request
by determining which device flag is set and then jump to a routine to
service the appropriate device. The service routine ends with a HLT,
which would be encountered only if the service routine is entered and
neither flag is set — a condition that should never exist.

The keyboard input routine is entered when the keyboard flag is set.
The flag is cleared to prevent further interruptions when the interrupt
system is re-enabled. If the mode is not set for input, program control
is transferred to the background program. Otherwise, the program
accepts the character (KRB), acknowledges its receipt by printing it
on the printer (TLS), and stores it in the buffer. (Notice that no KSF,
JMP . — 1 loop is necessary.) The routine then checks for the ASCII
code for a period, returning to the background program if it is not a
period. Upon receipt of a period, the routine resets the buffer and sets
the mode for output. Program control then returns to the background
program. (Since a TLS instruction was executed previously, an interrupt
will be requested when this action is complete, and the stored ASCII
codes will be typed out by the printer output routine.)

The printer output routine is entered when its flag is set. The routine
clears the device flag and checks for output mode. When in output
mode, the routine prints one character from the buffer. (Notice that no
TSF, JMP .—1 loop is necessary.) If the character is not a period,
control returns to the background program while the printer finishes
typing the character. If the character is a period, the routine resets the
buffer, sets the mode for input, and returns to the background program.

5-25

The exit routine to return to the background program must restore
the link and accumulator to the values at the time of interrupt. The
program turns the interrupt on (ION) and then returns to the rotate
program via a JMP I instruction (location contains the value of
the program counter when the interrupt occurred). The ION instruc-
tion does not take effect until the instruction following it has been
executed.

The constants used by the routines conclude the program listing.

RESERVE
BUFFER SPACE
FOR CHARACTERS

TURN
INTERRUPT ON

ROTATE BIT

ENDLESSLY THROUGH

ACCUMULATOR

SAVE AC
AND L

RESTORE
ftC AND L

RETURN TO

BACKGROUND

PROGRAM

KEYBOARD^
FLAG SET

? j*

S*ES

s. res

ENTER KEYBOARD

SERVICE ROUTINE

(KB)

[NO

PRINTERS-
FLAG SET
7 ^

ENTER PRINTER
SERVICE ROUTINE

(TP1

CLEAR
KEYBOARD FLAG

Figure 5-1 5A. Program to Operate on Program Interrupt Facility:

Flowchart

5-26

INE.

/FIRST INSTRUCTIONS

JMP I 2

SERV
*200

/INITIALIZATION ROUT
START, CLA CLL

DCA MODE

TAD Kl 777

DCA BUFFER

ION
/BACKGROUND PROGRAM
ROTATE, CLA CLL IAC

ISZ COUNT

JMP .-1

ISZ COUNT

JMP . -1

RAL

JMP ROTATE+1
/SERVICE ROUTINE
SERV, DCA AC

RAL

DCA L

KSF

SKP

JMP KB

TSF

SKP

JMP TP

HLT
/KEYBOARD INPUT ROUTINE

AFTER AN INTERRUPT.

/STORES RETURN ADDRESS.
/JUMP TO SERVICE ROUTINE.

KB, KCC

TAD MODE
SZA CLA
JMP EXIT
'ISZ BUFFER
KRB
TL.S'.

DCA I BUFFER
TAD I BUFFER
TAD MPER
SZA CLA
JMP EXIT
TAD K1777
DCA BUFFER
CLA CMA
DCA MODE
JMP EXIT

/SET MODE FOR INPUT.

/SET UP BUFFER SPACE.

/TURN ON INTERRUPT FACILITY.

/SET ONE BIT IN AC.
/ROTATING DELAY INSTRUCTIONS.

/ROTATE BIT LEFT.

/EXECUTE DELAY INSTRUCTIONS.

/SAVE ACCUMULATOR.

/SAVE LINK.

/KEYBOARD INTERRUPT?

/NO: CHECK PRINTER.

/YES: SERVICE KEYBOARD.

/PRINTER INTERRUPT?

/NO: SKIP PRINTER ROUTINE JMP.

/YES: SERVICE THE PRINTER.

/FATAL HALT IF NO FLAG SET.

/CLEAR KEYBOARD FLAG.
/INPUT MODE?

/NO: RETURN TO BACKGROUND.
/YES: INCREMENT BUFFER.
/READ THE CHARACTER.
/ACKNOWLEDGE ON THE PRINTER.
/STORE CHARACTER.
/IS CHARACTER A PERIOD?

/NO: RETURN TO BACKGROUND.

/YES: RESET BUFFER

/TO TYPE THE CHARACTERS.

/SET MODE FOR OUTPUT AND
/RETURN TO BACKGROUND.

Figure 5-15B. Program to Operate on Program Interrupt Facility: Coding

5-27

/PRINTER OUTPUT ROUTINE

rp,

TCF

TAD

MODE

SNA

CLA

JMP

EXIT

ISZ

BUFFER

TAD

I BUFFER

TLS

TAD

MPER

SZA

CLA

JMP

EXIT

DCA

MODE

TAD

Kl 777

DCA

BUFFER

JMP

EXIT

/ROUTINE FOR RETURNING

EXIT,

TAD

CLL

RAR

TAD

I ON

JMP

COUNT,

MODE,

K1777,

1 777

BUFFER,

AC,

MPER,

-256

/CLEAR PRINTER FLAG.
/OUTPUT MODE?

/NO: RETURN TO BACKGROUND.
/YES: INCREMENT BUFFER.
/GET CHARACTER FROM BUFFER.
/TYPE IT OUT.
/IS CHARACTER A PERIOD?

/NO: RETURN TO- BACKGROUND.
/YES: SET MODE FOR INPUT,
/RESET BUFFER, AND

/RETURN TO BACKGROUND.
TO BACKGROUND PROGRAM.
/RESTORE LINK.

/RESTORE ACCUMULATOR.
/RE-ENABLE INTERRUPT FACILITY.
/0 CONTAINS RETURN ADDRESS.

Figure 5-1 5B (cont.)

Program to Operate on Program Interrupt Facility:
Coding

Advanced Use of the Program Interrupt Facility

The following paragraphs entitled "Multiple Device Interrupt Pro-
gramming" and "A Software Priority Interrupt System" are intended
for programmers making extensive use of the program interrupt facil-
ity. The reader who merely desires a general knowledge of the interrupt
facility may omit these topics from his reading.

MULTIPLE DEVICE INTERRUPT PROGRAMMING

Many programming applications use the interrupt facility to service
several devices. For example, a PDP-8 may use the program interrupt

5-28

facility to control the operation of DECtape and DECdisk systems
through a Teletype console. Systems of this type require a service
routine that determines the source of the interrupt request (i.e., which
device flag is set). The following example is an instruction sequence
which uses dummy skip instructions to determine the device requesting
the interrupt.

DASF

SKP

JMP SERVA / DEVICE A REQUESTED THE INTERRUPT

DBSF

SKP

JMP SERVB /DEVICE B REQUESTED THE INTERRUPT

DCSF

SKP

JMP SERVC / DEVICE C REQUESTED THE INTERRUPT

DNSF

SKP

JMP SERVN / DEVICE N REQUESTED THE INTERRUPT

The dummy skip instructions (DASF, DBSF, etc.) are skip-on-flag
instructions for each of the devices in the interrupt system. (Usually,
these instructions skip the next instruction if the device flag is set to a
1. Instructions for some devices, however, may skip if the flag is a 0,
i.e., skip-on-non-flag. Instructions of this type should not be followed
by the unconditional SKP instruction.) Because of the predominance
of SKP instructions, the instruction sequence which determines the
source of an interrupt request is often called a skip chain.

As the previous example implies, the skip chain may be enlarged to
test for almost any number of device flags. However, an important
limitation upon the size of a skip chain is imposed by the devices which
the chain serves. The chain must be traversed and the device serviced
before the desired information is lost. High-speed devices, such as mag-
netic tapes, disks, or drums, must be serviced quickly or the informa-
tion will not be available. On the other hand, low-speed devices, such
as paper tape readers and punches, have a relatively long period of
time in which they may be serviced before any loss of information
occurs. The programmer must take these time factors into account
whenever he programs a multiple device interrupt system.

5-29

High-speed devices should be tested and serviced first by the skip
chain. Thus, the chain should begin by checking the device flags of the
high-speed devices of the system, such as DECtape and DECdisk, and
conclude by checking the flags of the low-speed devices such as the
Teletype keyboard or paper tape readers and punches. Thus, a high-
speed device is not required to wait while a long set of device flags
is checked.

In the case of two interrupt requests occurring simultaneously, the
high-speed device is serviced first. The second device may be serviced
simply by returning to the background program and waiting for an-
other request, or it may be serviced by checking the flags through a
skip chain before returning to the background program.

A SOFTWARE PRIORITY INTERRUPT SYSTEM

A service routine may be written in such a way that a priority of de-
vice interrupts is established through software. The programmer sets
priorities by allowing the service routine for a particular device to be
interrupted by the interrupt request of a higher priority device. This
may be necessary in a system that includes high-speed devices that re-
tain information for a short time and that require immediate attention.
A lower-priority device would be serviced by a routine that would re-
enable the interrupt facility at its beginning. The higher-priority device
would not re-enable the interrupt until it had completed its task.

The service routine of a multiple interrupt system must include in-
structions to save the contents of the accumulator and link for each
interrupt. The contents of the program counter must also be removed
from location before a new interrupt occurs. To save the contents of
these active registers, the programmer must establish a "push down"
list of accumulators, links, and program counters. The instructions to
handle such a list could be the following.

SERV, DCAIACPTR /SAVE AC
RAL

DCAILPTR /SAVEL
TADO

DCA I PCPTR /SAVE PROGRAM COUNTER

ISZACPTR /INCREMENT POINTERS FOR

ISZ LPTR /NEXT USE.

ISZ PCPTR

5-30

The above instructions would then be followed by the chain of instruc-
tions to test the flags and jump to the service routine for the device
whose flag is set.

The service routine for a high-priority device in such a system is
similar to the preceding program examples, in that the interrupt facility
is not enabled until processing is complete.

The low-priority device would be serviced by a routine which would
re-enable the interrupt system immediately after clearing its device
flag. A low-priority keyboard routine might start as follows.

KB, KCC /CLEAR DEVICE FLAG.

ION /TURN INTERRUPT ON.

TAD MODE / SERVICE PROGRAM WILL

/PROCEED UNTIL

/ INTERRUPTED OR COMPLETED.

Thus, the keyboard could be interrupted during the process of trans-
mitting a character. The service routine would save the status of the
routine by storing the program counter accumulator and link in the
push-down list described before. After the high priority device had
been serviced, the following exit routine would return control to the
low priority device.

EXIT, IOF /INTERRUPTS NOT ALLOWED.

CLACMA /FOLLOWING INSTRUCTIONS WILL

TAD ACPTR /DECR EASE THE POINTERS

DCA ACPTR /BY t AND THEN

CLA CMA /USE THEM.

TAD LPTR

DCA LPTR

CLACMA

TAD PCPTR

DCA PCPTR

TAD I PCPTR

DCAO

TAD I LPTR

CLL RAR

TAD I ACPTR

JMP I

Through this approach, multiple interrupts could occur and would be
serviced on a priority basis specified by the programmer.

5-31

Program Interrupt Demonstration Program

The program presented in Figures 5-16A through 5-16F is a dem-
onstration program to run on the program interrupt facility. It contains
a bit rotating program, the speed and direction of which is determined
by the switch register settings. (This same program is presented in
Exercise 10 of Chapter 3.) The foreground program is the ordering
program which was given in Figure 5-14. This program has the capac-
ity to accept 4-digit positive octal input from the Teletype keyboard,
automatically terminating each 4-digit number with a carriage return
and line feed. Upon receipt of a typed dollar sign ($) , the program will
place the data in increasing order, and type the ordered data on the
printer. The program will not accept negative numbers or nonoctal
digits.

The example is useful as a demonstration and illustration of the
power of program interrupt as the computer will seem to be perform-
ing two tasks at the same time. The programmer knows that this is not
possible and that the two tasks are sharing the computer time; how-
ever, the appearance indicates simultaneous actions.

When an interrupt request occurs, locations 0, 1, and 2 of page
provide the storage for the program counter and the jump to the service
routine.

The constants which are stored on page include four "software
switches" to record the conditions within the running program. MODE
is used to specify the input or output status of the running program.
SW1 is used to signal the input of the first digit of a new number (0)
or the input of successive digits of a continuing number ( 1 ) . SW2 is
used to control the input and output of data by separating each number
with a carriage return and line feed. SW3 is used to bypass the output
mode and allow the typing of carriage returns, line feeds, and question
marks (to denote errors of input) during the input of data.

Other constants include pointers which permit off-page jumps to
routines elsewhere in memory. The constant BUFF is a pointer to the
storage area for the packed input numbers. It contains END which is
defined as END = . in the last line of Figure 5-16F. Thus, the buffer
is all of memory following the last instruction.

5-32

/FIRST INSTRUCTIONS AFTER AN INTERRUPT.

JMP I 2

SERV
*50

/CONSTANTS STOR
MODE,

SW1,

SW2,

SW3,

AC,

PRINTR,

KEY8RD,

ORDPTR,

EXITPT,

M7000,

BUFF,

BUFFPT,

M4,

DIGCTR,

TEMP 1 ,

TEMP, .

MDOLAR,

M10,

HOLD,

HOLDL,

M260,

AMOUNT,

FLAG,

TALLY,

XI,

X2,

PRNTCT,

K7,

K260,

K212,

K215,

QUEST,

K4000,

ORDER

EXIT

1000

END

7774

• +2

7534
-10

•7520

260
212
215
277
4000

ED ON PAGE 0.
/INPUT=0J OUTPUT=-l.
/NUMBER STATUS SWITCH
/OUTPUT sCR=0,LF=-l,DATA=l .
/MODE BYPASS SWITCH
/SAVE AC AND
/L DURING AN INTERRUPT.
/FOLLOWING ARE POINTERS FOR
/THE RESPECTIVE ROUTINES.

/ORDER SUB-PROGRAM CONSTANTS

Figure 5-1 6A. Program Interrupt Demonstration Program
(Constants Located on Page 0)

5-33

/SUBROUTINES STORED ON

CRi

LFj

SW2SETj

TAD
TLS
CLA

DCA
JMP

K215

CMA

SW2

I EXITPT
TAD K212
TLS
CLA

TAD SW3
SNA CLA
JMP SW2SET
DCA SW3
DCA SW2
JMP I EXITPT
CLA CLL IAC
DCA SW2
JMP I EXITPT

PAGE 0.
/CARRIAGE RETURN ROUTINE

/SET SW2 FOR A LF.
/LINE FEED ROUTINE

/TURN OFF MODE BYPASS.
/SET SW2 FOR CR •

/SET SW2 FOR DATA,

Figure 5-1 6B. Program Interrupt Demonstration Program
(Subroutines Located on Page 0)

The location TEMP is used as a pointer to the four locations after
it; these locations are used for the storage of incoming ASCII codes.

The subroutines CR and LF type the carriage returns and line feeds
when called for by the setting of SW2.

The program begins in location 200 as shown in Figure 5-16C. The
initialization routine sets each of the software switches to zero. After
the initialization is completed, the interrupt facility is turned on and the
background program is started.

The rotate subprogram begins by checking the setting of the switch
register to determine the direction of rotation. The value of bit
specifies a rotate right when it is a 0, or a rotate left when it is a 1. The
last nine bits of the switch register determine the speed of the rotation.
They are stored in COUNT and determine the number of passes
through the ISZ COUNTR, JMP .-1 loop.

The BEGIN routine determines the speed and direction and the GO
routine establishes the bit position after each check of the switch regis-
ter setting. The INSTR routine executes the delay and the rotation of
the bit.

5-34

*200

/NEXT INSTRUCTIONS INITIALIZE THE PROGRAM.

/FURTHER INITIALIZATION DONE BY RESTART.

START,

I OF

/INTERRUPT OFF DURING INITIALI

CLA

CLL

DCA

MODE

DCA

SWI

DCA

SW2

DCA

SW3

TAD

BUFF

DCA

BUFFPT

DCA

AMOUNT

ION

/ROTATE

SUB-

-PROGRAM

BEGINS HERE.

ROTATE,

CLA

CLL CML

BEGIN,

DCA
RAL

SAVEAC

DCA

SAVEL

TAD

K7000

/ALWAYS SET BITS 0, 1 A

OSR

DCA

COUNT

/MAX COUNT IS -1000.

OSR

RAL

/PUT BIT IN LINK.

SZL

CLA

JMS

LEFT

JMS

RIGHT

CLL

GO,

TAD
RAR

SAVEL

TAD

SAVEAC

INSTR,

HLT

/OVERWRITTEN BY RAR OR

ISZ

COUNTR

JMP

INSTR+1

ISZ

COUNT

JMP

INSTR+1

JMP

BEGIN

SAVEAC,

SAVEL,

K7000,

7000

COUNTR,

COUNT,

/SUBROUTINES TO DETERMINE DIRECTION.

LEFT,

ISZ

LEFT

/SKIP INSTR AFTER JMS 1

TAD

KRAL

DCA

INSTR

JMP

I LEFT

/STORE 'RAL' IN 'INSTR

RIGHT,

TAD

KRAR

DCA

INSTR

/STORE "RAR" IN 'INSTR

JMP

I RIGHT

KRAR,

RAR

KRAL,

RAL

RAL.

LEFT.

Figure 5-16C. Program Interrupt Demonstration Program
(Initialization Routine and Rotate Subprogram)

5-35

/SKIP CHAIN TO SERVICE ROUTINES

SERV, DCA

RAL

DCA

TSF

SKP

JMP

I PRINTR

KSF

SKP

JMP

I KEYBRD

HLT

/ORDER SUB-PROGRAM

ORDER. CLA

CLL

TAD

AMOUNT

CIA

I AC

DCA

TALLY

DCA

FLAG

TAD

BUFF

DCA

TAD

BUFF

IAC

DCA

TEST, TAD

I X2

CIA

TAD

I XI

SPA

SNA CLA

JMP

INCPTR

REVERSE, TAD

I XI

DCA

HOLD

TAD

I X2

DCA

1 XI

TAD

HOLD

DCA

I XS

CLA

CLL CMA

DCA

FLAG

INCPTR, ISZ

ISZ

TALLY

JMP

TEST

TAD

FLAG

SZA

CLA

JMP

ORDER

CLA

CMA

DCA

MODE

TAD

BUFF

DCA

BUFFPT

DCA

SW1

TAD

AMOUNT

CIA

DCA

PRNTCT

TLS

JMP

I EXITPT

/SAVE AC AND L
/DURING AN INTERRUPT.

/IS INTERRUPT CAUSED
/BY TELETYPE PRINTER?
/SERVICE TELETYPE PRINTER.
/IS INTERRUPT CAUSED
/BY KEYBOARD?
/SERVICE KEYBOARD.
/SHOULD NEVER REACH HERE.

/SET TALLY FOR COMPARISONS.

/CLEARS FLAG FOR EACH PASS

/IS X2 LESS THAN XI ?

/NO: DON'T REVERSE.
/YES: REVERSE X2 AND XL

/SET FLAG TO SI GNAL
/THAT A REVERSE WAS DONE
/INCREMENT X POINTERS.

/COMPARED ALL ENTRIES?
/NO: COMPARE NEXT X'S.
/YES: ORDER DONE YET?

/NO: MAKE ANOTHER PASS.

/YES: SET OUTPUT MODE.

/SET POINTER TO FIRST ENTRY.

/CLEAR NUMBER STATUS SWITCH.
/SET A TALLY FOR OUTPUT.

/TO TRIGGER NEXT INTERRUPT.

Figure 5-1 6D. Program Interrupt Demonstration Program
(Skip Chain to Service Routines and Order Subprogram)

5-36

The service routines are reached through the skip chain in Figure
5-16D. These service routines are located on a separate memory page
and are reached through pointer words stored on page 0.

The order subprogram is initiated when input is complete. The
routine sorts the positive octal numbers stored in the buffer into in-
creasing order. The technique is the same as that used previously in the
program of Figure 5-14.

These routines conclude the instructions placed on page 1 of mem-
ory. The service routines begin on the next memory page.

*400

KB*

KCC

TAD

MODE

SZA

CLA ■

JMP

EXIT

TAD

SW1

SZA

CLA

JMP

CNTDI G

TAD

DCA

DIGCTR

TAD

TEMPI

DCA

TEMP

CNTDIG*

KRS
TLS

DCA

I TEMP

CHECK*

TAD

I TEMP

TAD

MDOLAR

SNA

CLA

JMP

I ORDPTR

TAD

I TEMP

TAD

M260

SPA

JMP

ERROR

TAD

M10

SPA

JMP

LEGAL

ERROR,.

CLA

I.AC

DCA

SW3

. DCA

SW1

TLS

/INPUT MODE DOES NOT
/HONOR KEYBOARD REQUEST.

/CHECK FOR A NEW NUMBER
/OR A CONTINUED DIGIT.

/READ KEYBOARD CHARACTER.
/TYPE IT ON PRINTER.
/STORE DIGIT TEMPORARILY.

/CHECK FOR TERMINAL $

/ASCII LESS THAN 260?

/YES: ERROR.
/NO: SUBTRACT 10.
/GREATER THAN 267?
/NO: DIGIT IS LEGAL.
/NOT AN OCTAL NUMBER.
/SET TO TYPE ?*CR*LF.
/SET FOR A NEW NUMBER.

Figure 5-1 6E. Program Interrupt Demonstration Program
(Keyboard Service Routine)

5-37

The service routines begin in location 400 as presented in Figure
5-16E. The keyboard service routine honors an interrupt only if the
mode is set to 0. The routine uses SW1 to specify number status, re-
setting the digit counter if a new number is started. The routine accepts
the incoming digit and types it on the printer.

The input of a non-octal character or terminal symbol ($) is checked
after each character is received. Whenever the terminal $ is received
by the program, control is transferred to the qrder subprogram. If the
character is not an octal digit, the program types a question mark and

JMP

EXIT

LEGAL,

CLA

CMA

DCA

SW1

ISZ

TEMP

ISZ

DIGCTR

JMP

EXIT

PACK,

TAD

TEMPI

DCA

TEMP

DCA

HOLD

TAD

DCA

DIGCTR

DIGPCK,

TAD

HOLD

RAL

CLL

RTL

TAD

I TEMP

TAD

M260

DCA

HOLD

ISZ

TEMP

ISZ

DIGCTR

JMP

DIGPCK

TAD

HOLD

DCA

I BUFFPT

TAD

I BUFFPT

TAD

K4000

SPA

CLA

JMP

NOTNEG

I AC

JMP

DISALO

NOTNEG,

ISZ

BUFFPT

ISZ

AMOUNT

CLA

CMA

DISALO,

DCA

SW3

DCA

SW1

TLS

JMP

EXIT

/SET SW1 TO SIGNAL
/A CONTINUED NUMBER.

/HAVE A DIGITS?

/NO: GET NEXT DIGIT.

/YES: PUT NUMBER TOGETHER.

/NEXT 7 INSTRUCTIONS
/COMBINE THE A OCTAL DIGITS
/INTO ONE MEMORY WORD.

/PACK ANOTHER DIGIT.

/STORE THE OCTAL NUMBER.
/CHECK FOR NEGATIVE ENTRY.

/ENTRY IS LEGAL.

/SET SW3 TO TYPE A "?".

/DISALLOW NEGATIVE ENTRY.

/COUNT THE ENTRIES.
/TYPE A CR,LF
/AFTER THE ENTRY.
/CLEAR SW1 FOR NEXT PASS.

Figure 5-16E (cont.). Program Interrupt Demonstration Program
(Keyboard Service Routine)

5-38

ignores the whole entry in which it occurred. The program requests a
new 4-digit number by typing a carriage return and line feed.

When four digits have been received* the pack routine combines the
four ASCII codes into one octal number and stores the number in the
buffer. If the number is negative, the entry is disallowed. (The next
packed number will be deposited in the same location, thus destroying
the negative number.) A running count of the entries is kept and later
used when ordering is performed. The software switches are finally set
to return the carriage for the next number to be input.

The printer service routine in Figure 5-16F is used to output the
ordered numbers and to type carriage returns, line feeds and question
marks during input. The "mode bypass switch", SW3, is used in con-
junction with the subroutines on page to type carriage returns and
line feeds. If the program is in output mode, the results of the sorting
will be typed by the printer service routine. The stored numbers will be
unpacked, translated into ASCII codes and typed out.

/PRINTER SERVICE ROUTINE

TP.

MODCHKj

RETLF,

DATA,

, TCF
TAD
SNA
JMP
TAD
SPA
JMP
TAD
TLS
CLA
DCA
JMP
f'AD
SNA
JMP
TAD
SNA
JMP
TAD
SPA
JMP
TAD
SZA
JMP
TAD
DCA
DCA
DCA

SW3

CLA'

MODCHK

SW3

CLA

RETLF

QUEST

CMA

SW3

EXIT

MODE

CLA

EXIT

SW2

CLA

SW2

CLA

SW1

CLA

DI GTYP

DIGCTR

HOLD

HOLDL

/CHECK MODE-BYPASS-SWITCH,
/NO' BYPASS, CHECK MODE.

/MODE-BYPASS,
/MODE BY PASS

DO A CR
SET FOR

& LF.

/AFTER TYPING THE ?
/TO TYPE THE CR &LF

/CHECK MODE.

/INPUT MODE:
/OUTPUT MODE:

IGNORE
BEGIN

SET SW3.

REQUEST!
OUTPUT.

/FIRST PASS, CARRIAGE RETURN.

/LINE FEED ON SECOND PASS.
/PRINT DATA ON THIRD PASS.
/NEW NUMBER?

/NO: TYPE ANOTHER DIGIT.
/YES: RESET DIGIT COUNTER.

/CLEAR THESE LOCATIONS.

Figure 5-16F. Program Interrupt Demonstration Program
(Printer Service, Exit and Restart Routines)

5-39

When the action of either service routine is completed, the exit rou-
tine returns control to the background program until the next interrupt
occurs. This routine restores the accumulator and link and turns the
interrupt on before jumping to the interrupted background program.

The restart routine initializes the software switches for the ordering
of a new set of data input from the keyboard. The mode is set for input
and control returns to the background program until new input is
suppie . ._ _,_„, /GET NUMBER T0 BE printed.

/ROTATE INTO THE LINK.
/THESE TWELVE INSTRUCTIONS
/PRINT OUT THE NEXT DIGIT.

TAD

I BUFFPT

DIGTYP, TAD

HOLDL

CLL

RAL

TAD

HOLD

RAL

RTL

DCA

HOLD

RAR

DCA

HOLDL

TAD

HOLD

AND

TAD

K260

TLS

CLA

CMA

DCA

SW1

ISZ

DIGCTR

JMP

EXIT

CLA

DCA

SW1

DCA

SW2

ISZ

BUFFPT

ISZ

PRNTCT

JMP

EXIT

RESTART, CLA

CLL

DCA

MODE

DCA

SW1

DCA

SW2

CMA

DCA

SW3

TAD

BUFF

DCA

BUFFPT

DCA

AMOUNT

JMP

EXIT

/ROUTINE TO

RETURN '

EXIT, CLA

CLL

TAD

/SET FOR ANOTHER DIGIT.

/WORD COMPLETE?

/NO: GET ANOTHER DIGIT.

/YES: SIGNAL A NEW NUMBER.

/TYPE A CR & LF.

/ALL NUMBERS PRINTED?
/NO: WAIT FOR NEW NUMBER.
/YES: SET UP FOR NEW INPUT.

/SET BYPASS SWITCH
/TO TYPE A CR & LF.

TO ROTATE PROGRAM
CLL
L
RAR

TAD AC
ION

JMP I (
END=.
s • „

Figure 5-16F (cont.). Program Interrupt Demonstration Program
(Printer Service, Exit and Restart Routines)

5-40

/0 CONTAINS RETURN ADDRESS.

DATA BREAK

Programmed transfers of data, including program interrupt transfers,
pass through the accumulator. The accumulator must therefore be
cleared while the transfer is performed. This type of transfer is often
too slow for use with extremely fast peripheral devices. Devices which
operate at very high speed, or which require very rapid response from
the computer, use the data break facility (standard on all PDP-8
family computers except the PDP-8/L). Use of these facilities permits
an external device to insert or extract words from the computer mem-
ory, bypassing all program control. Because the computer program has
no cognizance of the transfers made in this manner, the program must
check for the presence of this data prior to its use. The data break is
particularly well-suited for devices that transfer large amounts of data
in block form, for example, random access disk files, high-speed mag-
netic tape systems, or high-speed drum memories.

The data break facility allows a peripheral device to transfer infor-
mation directly with the PDP-8 core memory on a "cycle stealing"
basis. Input/output equipment operating at high speed can transfer in-
formation with the computer through the data break facility more
efficiently than through programmed means. In contrast to programmed
operations, the data break facilities permit an external device to con-
trol information transfers.

Data breaks are of two basic types: single-cycle and 3-cycle. In a
single-cycle data break, registers in the device specify the core memory
address of each transfer and count the number of transfers to deter-
mine the end of the "block. In the 3-cycle data break, two computer
memory locations perform these functions, simplifying the device in-
terface by omitting the two hardware registers.

In general terms, to initiate a data break transfer of information, the
control must do the following tasks.

1 . Specify the affected address in core memory.

2. Provide the data word.

3. Indicate direction of data word transfer.

4. Indicate a single-cycle or 3-cycle break.

5. Request the data break.

5-41

Single-cycle data break Is a device-controlled transfer of information
which steals one memory cycle from the PDP-8. When a device that is
connected to the data break facility wishes to transfer a word of infor-
mation, it requests a data break. The PDP-8 computer completes the
current instruction and enters the break state. For one memory cycle,
the device has access to the PDP-8 and transfers the data word into
(or out of) the memory unit at a location specified by the device con-
trol. The device controls the number of words to be transferred and
the locations in memory to be affected by the transfer. The device con-
tinues to request breaks and transfers one word per break until the
block transfer is complete.

The 3-cycle data break facility provides a current address register
and a word count register in core memory for each connected device,
thus eliminating the necessity for registers in the device control. When
several devices are connected to the facility, each is assigned a different
pair of core locations for word count and current address, allowing
interlaced operations of the devices. The device specifies the location
of these registers in memory. Since these instructions are in memory,
they may be loaded and unloaded without using IOT instructions.

The 3-cycle data break facility performs the following sequence of
operations.

1. An address is read from the device to indicate the location of
the word count register. This location specifies the number
of words in the block yet to be transferred. The address is
always the same for a given device.

2. The content of the specified word count register is read from
memory and is incremented by 1. To transfer a block of n
words, the word count is set to — n during the programmed
initialization of the device. When this register is incremented
to 0, a pulse is sent to the device to terminate the transfer.

3. The location after the word count register contains the cur-
rent address register for the device transfer. The content of
this register is set to 1 less than the locati&n to be affected by
the next transfer. To transfer a block beginning at location
A, the register is originally set to A-l.

4. The content of the current address register is incremented by
1 and then used to specify the location affected by the trans-
fer.

5-42

After the transfer of information has been accomplished through the
data break facility, input data (or new output data) is processed,
usually through the program interrupt facility. An interrupt is re-
quested when the data transfer is completed and the service routine
will process the information.

EXERCISES

1. Write a Subroutine ALARM which rings the teleprinter bell
five times.

2. Write a format subroutine for the teleprinter to tab space the
teleprinter carriage. The subroutine is entered with the num-
ber of spaces to be tabbed in the accumulator.

3. Write a program that will type a heading at the top of the
paper and then type the numbers 1 through 20 down the left
hand side of the page with a period after each number.

4. Write a program which will accept a 2-digit octal number
from the Teletype keyboard and type "SQUARED=" and
the value for the number squared, followed by "OCTAL"
and a carriage return and line feed.

5. Extend the program written in Exercise 4 by adding routines
to disallow the input of an 8 or 9 and type out an appropriate
message.

6. Combine the program of Exercise 5 with a bit-rotating pro-
gram to use the program interrupt facility.

5-43

NOTE TO READER

This chapter concludes the introduction to machine-language pro-
gramming on the PDP-8. The remaining chapters deal primarily with
software supplied by Digital Equipment Corporation as an aid to the
programmer. Chapter 6 describes the software available to assist the
PDP-8 programmer in writing machine-language programs. Later
chapters describe specialized software systems and a conversational
language, FOCAL, which can be used to write programs on the PDP-8.

The programming of advanced input/output devices and options
(e. g., EAE, DECtape, DECdisk, A/D Converters) are beyond the
scope of this publication. The applicable PDP-8 Users Handbook
should be consulted when programming these devices.

5-44

Chapter 6

Operating the
System Software

This chapter contains brief descriptions of the PDP-8 family systems
programs and selected program operating procedures. The major por-
tion of the chapter is devoted to detailed operating procedures for the
most frequently used system software — the loaders, symbolic editor, the
assemblers, the dynamic debugging programs, and FOCAL.

DESCRIPTIONS

A comprehensive package of system software accompanies each com-
puter in the PDP-8 family. (Software is the collection of programs and
routines associated with the computer.) The package contains many
programs and routines furnished on punched paper tape or stored on
DECtape in binary coded format, as well as associated manuals and
documents describing the use and operation of each program and rou-
tine.

The system software, supplied by Digital Equipment Corporation,
allows the programmer to write, edit, assemble, compile, debug, and
run his programs, making the full data processing capability of the
computer immediately available. System software comes from past,
present, and continuing programming efforts of DEC programmers and
, users (see Chapter 11).

The system programs furnished with each computer in the PDP-8
family are those capable of operating with that 4 specific computer and
its I/O devices. If, for example, a computer has 4K of core memory,
the set of programs in the accompanying software package are those
designed to run in 4K of core memory, and the package accompanying
computers with 8K of core memory include those programs capable of
operating in 8K of core memory, i.e., both 4K and 8K programs, and
so on.

Many of the frequently used software programs are briefly described
on the following pages.

6-1

Symbolic Editor

The Symbolic Editor is a program which allows the programmer to
prepare, edit, and generate a symbolic program tape online from the
Teletype keyboard. Using Editor, the programmer may enter his sym-
bolic program into core from the keyboard or paper tape reader, and
then issue certain commands to edit his program. When used properly,
Editor can substantially ease the labor of writing and editing symbolic
programs and reduce the number of passes necessary to correct sym-
bolic program tapes. Symbolic Editor is further described later in this
chapter and in the DEC manual entitled Symbolic Editor, Order No.
DEC-08-ESAB-D.

PAL III Symbolic Assembler

The PAL III Symbolic Assembler is a two-pass assembler which
translates symbolic programs written in the PAL III symbolic language
into binary-coded programs, producing the binary tape acceptable to
the computer. The assembler offers an optional third pass which pro-
duces an octal/symbolic printout and/or punchout of the assembled
program. The assembler is described in more detail later in this chapter
and in PAL III Symbolic Assembler, Order No. DEC-08-ASAC-D.

MACRO-8 Symbolic Assembler

The MACRO-8 Symbolic Assembler accepts symbolic programs
written in the MACRO-8 symbolic language and translates them into
binary-coded programs in two passes. An optional third pass is avail-
able for an octal/symbolic assembly program listing. MACRO-8 is
compatible with PAL III and has the following alditional features:
user-defined macros, double-precision integers, floating-point constants,
•Bthmetic and Boolean operators, literals, text facilities, and automatic
off-page linkage generation. The MACRO-8 Symbolic Assembler is
discussed later in this chapter and in MACROS Assembler, Order No.
DEC-08-CMAA-D.

8K SABR Symbolic Assembler

SABR is an advanced one-pass assembler for use with 8K to 32K
words of core. The SABR language is similar to the assemblers above
with many additional features, and differs from them in its operating
procedures, pseudo-ops, assembled output (relocatable binary code),
and execution of assembled programs. SABR is also used with 8K
FORTRAN. For a complete description, refer to 8K SABR Assembler,
Order No. DEC-08-ARXA-D.
FORTRAN Compilers and Operating Systems

FORTRAN (FORmula TRANslation) is a problem-oriented lan-
guage written mostly in mathematical terms and some English words. It

6-2

is especially suited for solving equations and making other mathematical
calculations,. For those who elect to write their programs in FORTRAN,
DEC has 4K FORTRAN and 8K FORTRAN. Each is briefly ex-
plained below.

4K FORTRAN

4K FORTRAN consists of a compiler, debugging aid, and oper-
ating system. The one-pass compiler translates FORTRAN symbolic-
language statements into binary code and produces a binary tape. The
debugging aid (Symbol Print) lists the variables used and their locations
in core and indicates the section of core used by the compiled program.
The operating system loads and executes the compiled program. In ad-
dition, the operating system contains an extensive library of arithmetic
function subprograms and I/O routines. Useful error messages are
printed on the teleprinter when any error is detected by the compiler or
operating system. For additional information, refer to 4K FORTRAN,
Order No. DEC-08-AFC0-D.

8K FORTRAN

8K FORTRAN consists of a one-pass compiler, the one-pass 8K
SABR assembler, a linking loader, and an operating system, as well as
a comprehensive library of subprograms. During compilation, the sym-
bolic program is compiled into nonexecutable binary code. During
assembly, a relocatable binary program tape is produced. The linking
loader is used to convert the relocatable binary code into absolute
binary code for execution under control of the operating system.
Meaningful error messages are typed on the teleprinter as they are
detected. For additional information, refer to 8K FORTRAN, Order
No. DEC-08-KFXB-D.

ALGOL-8

ALGOL (for ALGOrithmic Language) is one of the most widely
used international programming languages. The language emphasizes
formal, well-defined procedures for solving problems with computers
and is the standard language of the scholarly Association for Computing
Machinery (ACM).

The ALGOL-8 compiler conforms to SUBSET ALGOL 60 as
approved by the International Federation of Information Processing
Societies (IFIPS) with additional restrictions. Further information is
provided in ALGOLS Programmer's Reference Manual, Order No.
DEC-08-KAYA-D.

6-3

FOCAL

FOCAL (for FOrmula CALculator) is an online, conversational in-
terpreter designed to be used as a tool by students, engineers, and scien-
tists in solving a wide variety of their problems. The language consists
of short imperative English statements and mathematical expressions in
standard notation. FOCAL puts the full calculating power and speed of
the computer at the user's fingertips without the user having to master
the intricacies of machine-language programming; in fact, the user need
know nothing at all about computers.

Using FOCAL, a FORTRAN-like program can be entered from the
keyboard and immediately executed, with the interpretive features
taking care of editing, compiling, and executing the stored program.

Procedures for loading and getting "online" with FOCAL are de-
scribed later in this chapter, and a thorough description of the FOCAL
language is given in Chapter 9. (FOCAL is also described in a separate
manual, Order No. DEC-08-A J AD-D. )

BASIC-8

BASIC-8 is a modified version of the very popular algebraic language
developed at Dartmouth College. The BASIC-8 language is composed
of easy-to-learn English statements and mathematical expressions. It Is
ideally suited for the classroom as well as the office or laboratory.
(BASIC-8 operates with TSS/8, described in Chapter 8.)

Disk Monitor System

The Disk Monitor System is a keyboard-oriented system con-
taining a monitor and a comprehensive software package. The pack-
age includes a FORTRAN Compiler, Program Assembly Language
(PAL-D), Editor program (Editor), Peripheral Interchange Program
(PIP), and Dynamic Debugging Technique (DDT-D) program. These
system programs simplify the user's task of editing, assembling, com-
piling, debugging, loading, saving, calling and running his own pro-
grams. The system is modular and open ended, permitting the user to
construct the software required in his particular environment and to
have full access to his disk for storage and retrieval of his programs.

Chapter 7 is devoted to the Disk Monitor System, and it is further
described in Disk Monitor System, Order No. DEC-D8-SDAB-D.

TSS/8

TSS/8 (Time-Sharing System for the PDP-8/I and PDP-8 com-
puters) is a general-purpose, stand-alone, time-sharing system. TSS/8

6-4

offers each of up to 16 users a comprehensive library of programs which
provide facilities for compiling, assembling, editing, loading, saving,
calling, debugging, and running user programs online. Any of these
library programs can be called into use by typing, in response to Moni-
tor's invitation (the dot), the command R and the assigned name of the
program. For example, .R FOCAL brings the FOCAL program into
core from the disk and automatically executes FOCAL so that it begins
"typing out its initial dialogue (see Chapters 8 and 9).

The heart of this time-sharing system is a complex of programs
called Monitor. Monitor coordinates the operations of the various units,
allocates the time and services of the computer to users, and controls
their access to the System. By segregating the central processing opera-
tions from the time-consuming interactions with the human users, the
computer can in effect work on a number of programs simultaneously.
The executions of various programs are interspersed without interfering
with one another and without detectable delays in the responses to the
individual users. See Chapter 8 or Time-Sharing System — TSS/8 Moni-
tor, Order No. DEC-T8-MRFB-D.

Loaders

A loader is a short program or routine which, when in core, enables
the computer to accept and store in core other programs. DEC offers
the programmer the following assortment of loaders to use, depending
on his preference and system configuration.

Read-In Mode (RIM) Loader — used to load into core programs
punched on paper tape in RIM format, primarily the Binary
Loader. RIM consist of 17 instructions which are toggled into core
using the console switches.

Binary (BIN) Loader — used to load into core programs punched
on paper tape in binary format, which includes the programmer's
binary tapes and most of DEC's system software. BIN is on
punched paper tape in RIM format.

HELP Loader — used to load into core the RIM and BIN Loaders.
HELP is in two parts: the first part consists of 11 instructions
which are toggled into core using the" console switches; the second
part is the HELP Bootstrap Loader punched on paper tape (con-
taining the RIM and BIN Loaders), which is loaded into core
using the low-speed paper tape reader.

TC01 Bootstrap Loader— used to load into core the DECtape
Library System programs. The loader is a 20-instruction program
which can be toggled into core using the console switches or it may

6-5

be on paper tape ia RIM format and therefore read into core using
the RIM Loader.

The four loaders above are covered in greater detail in the System
User's Guide, Order No. DEC-08-NGCB-D. RIM and BIN are also de-
scribed under "Operating Procedures" in this chapter.

Disk System Binary Loader — is a keyboard-oriented loader used
to input assembled binary programs into core. This loader is ex-'
plained in Chapter 7 and in Disk Monitor System, Order No.
DEC-D8-SDAB-D.

Dynamic Debugging Programs

Dynamic debugging programs are service programs that allow the
programmer to run his binary program on the computer. From the
Teletype keyboard, the programmer can control program execution,
examine registers and change their contents, make alterations to the pro-
gram, and much more. DDT-8 and ODT-8 are the two dynamic debug-
ging programs included in the system software package.

DDT-8 (Dynamic Debugging Technique) — allows the programmer
to do all the things mentioned in the preceding paragraph by com-
municating with his object program using either the mnemonic
coding of the symbolic program or the octal coding of the binary
program. DDT-8 is described in more detail later in this chapter
and in DDT-8, Order No. DEC-08-CDDA-D.
ODT-8 (Octal Debugging Technique) — allows the programmer to
do all the things mentioned above by communicating with his
object program using the octal representation of his binary pro-
gram. ODT-8 occupies less core storage than DDT-8 and can be
loaded in upper memory or lower memory, depending on where
the binary program resides. If the programmer's program uses
floating-point numbers, the low version of ODT-8 must be used
when debugging his program (DDT-8 does not interpret floating-
point numbers). ODT-8 is described in more detail later in this
chapter and in ODT-8 Order No. DEC-08-COCO-D.

Library of Utility Subroutines

Utility subroutines are short routines for performing such tasks as
printout or punchout of core memory content in octal, decimal, or
binary form, as specified by the programmer. Other tasks include octal
or decimal data transfers and binary-to-decimal, decimal-to-binary, and
paper tape conversions. These subroutines can be incorporated in the
user's program or run independently.

6-6

A complete set of standard diagnostic programs is provided to sim-
plify and expedite system maintenance; these are called MAINDEC
programs or routines. MAINDECs permit the programmer to effec-
tively test the operation of the computer for proper core memory func-
tioning and proper execution of instructions. They also enable perform-
ance checking of standard and optional peripheral devices. A list of
software documentation may be obtained from the DEC Program
Library.

Library of Mathematical Subroutines

The system software package also includes a set of mathematical
function routines to perform the following operations in both single and
double precision: addition, subtraction, multiplication, division, square
root, sine, cosine, arctangent, natural logarithm, and exponentiation.
These routines are incorporated in the programmer's symbolic program
as needed and are executed in response to an indirect JMS instruction
to the desired routine. (See Program Library Math Routines, Order No.
DEC-08-FFAC-D.)

Also included in the software package is a floating-point system to
enable the programmer to concentrate on the logic of his computation
rather than on decimal points. The system maintains a constant number
of significant digits throughout the computation, thereby enhancing the
accuracy of the result.

Floating-point notation is particularly useful for computations in-
volving numerous multiplications and divisions where magnitudes are
likely to vary widely and where only crude predictions can be made as
to the amount of variation involved. The floating-point system allows
storage of very large or very small numbers by storing only the signifi-
cant digits together with the exponent for that number. The system is
constructed as a self-contained package which includes its own input,
arithmetic, and output routines. It allows the programmer to use
floating-point arithmetic without having to construct his own arithmetic
subroutines.

When loaded in core, the starting address of the floating-point system
is stored in absolute location 0007 and is activated in response to a
JMS I 7 instruction. Then, the appropriate floating-point subroutines
are called by the subsequent instructions in the program, as shown in
the following example.

JMS I 7 /indirect jump to floating-point system,

FGET A /gets the floating-point number A

FADD B /and adds it to floating-point number B

6-7

FPUT C /and puts the result in floating-point

FEXT /location C, and then exits the floating-

/ point system, returning to the main

/program

For more detail, see Floating-Point System, Order No. Digital 8-5-S.

Availability of Software

System program tapes and manuals are available from DEC Sales
Offices and the central Program Library. In addition, a large variety of
programs written by users of PDP-8 family computers are available
through DECUS (see Chapter 11).

OPERATING PROCEDURES

Initializing the System

Before using the computer system, it is good practice to initialize all
units. To initialize the system, ensure that all switches and controls are
as specified below.

1. Main power cord is properly plugged in.

2. Teletype is turned OFF.

3. Low-speed punch is OFF.

4. Low-speed reader is set to FREE:

5. Computer POWER key is ON.

6. PANEL LOCK is unlocked.

7. Console switches are set to

DF=000 IF=000 SR=0000

SING STEP and SING INST are not set.

8. High-speed punch is OFF.

9. DECtape REMOTE lamps OFF.

The system is now initialized and ready for your use.

Loaders

READ-IN MODE (RIM) LOADER

When a computer in the PDP-8 family is first received, it is nothing
more than a piece of hardware; its core memory is completely demag-
netized. The computer "knows" absolutely nothing, not even how to

6-8

receive input. However, the programmer knows from Chapter 4 that he
can manually load data directly into core using the console switches.

The RIM Loader is the very first program loaded into the computer,
and it is loaded by the programmer using the console switches. The RIM
Loader instructs the computer to receive and store, in core, data
punched on paper tape in RIM coded format (see Chapter 4). (RIM
Loader is used to load the BIN Loader described below.)

There are two RIM loader programs: one is used when the input is
to be from the low-speed paper tape reader, and the other is used when
input is to be from the high-speed paper tape reader. The locations and
corresponding instructions for both loaders are listed in Table 6-1.

The procedure for loading (toggling) the RIM Loader into core is
illustrated in Figure 6-1 ,

Table 6-1. RIM Loader Programs

Instruction

Location

Low-Speed Reader

High-Speed Reader

7756

6032

6014

7757

6031

6011

7760 .

5357

7761

6036

6016

7762

7106

7763

7006

7764

7510

7765

5357

5374

7766

7006

7767

6031

6011

7770

5367

7771

6034

6016

7772

7420

7773

3776

7774

3376

7775

5356

5357

7776

0000

After RIM has been loaded, it is good programming practice to
verify that all instructions were stored properly. This can be done by
performing the steps illustrated in Figure 6-2, which also shows how to
correct an incorrectly stored instruction. *

When loaded, the RIM Loader occupies absolute locations 7756
through 7776.

6-9

Set SR»7756

Set SR ■ First
Instruction

Depress DEP

Set SR'Next
Instruction

Depress DEP

RIM Is Loaded

Set *
DF*Oesired Field
IF*Desired Field

* DECtape users should
load RIM into field 0.

Figure 6-1. Loading the RIM Loader

6-10

(

Initialize

)

^^ Using ^^_

Extended
* Memory ' .
\. ? ^r

Tno

.Yes

Set
OF » Correct Field
IF = Correct Field

Set SR ■ 7756

J-

Depress LOAD ADD

Depress EXAM

<y mb= ^.

Instruction
V ? s

Yes

Set SR = MA

Depress LOAD ADD

y^ All ^-v.

Instructions
^ Checked .
\. ? ^^

TVes

Set SR=Correct
Instruction

. RIM Is Loaded

Figure 6-2. Checking the RIM Loader

BINARY (BIN) LOADER

The BIN Loader is a short utility program which, when in core, in-
structs the computer to read binary-coded data punched on paper tape
and store it in core memory. BIN is used primarily to load the programs
furnished in the software package (excluding the loaders and certain
subroutines) and the programmer's binary tapes.

BIN is furnished to the programmer on punched paper tape in RIM-
coded format. Therefore, RIM must be in core before BIN can be
loaded. Figure 6-3 illustrates the steps necessary to properly load BIN.
And when loading, the input device (low- or high-speed reader) must
be that which was selected when loading RIM.

6-11

o— c

V ----- - - | See Figures 6-1 ,6-2

s^ Using ^v.
S*^ Extended ^n.
\ Memory v-^

Yes

Set*
DF = Correct Field
I F » Correct Field

|No

Set SR=7?56

Depress LOAD ADD

Htgh-Speed Reoder

Put BIN Tope
In HSR

Low -Speed Reoder

Turn TTY To LINE

Depress START

Put LSR To FREE

Rut SIN Tope
In LSR

Put LSR To START

Depress START

Figure 6-3. Loading the BIN Loader
6-12

When stored in core, BIN resides on the last page of core, occupying
absolute locations 7625 through 7752 and 7777.

BIN was purposely placed on the last page of core so that it would
always be available for use— the programs in DEC's software package
do not use the last page of core (excluding the Disk Monitor, dis-
cussed in Chapter 7). The programmer must be aware that if he writes
a program which uses the last page of core, BIN will be wiped out
when that program runs on the computer. When this happens, the pro-
grammer must load RIM and then BIN before he can load another
binary tape.

Figure 6-4 illustrates the procedure for loading binary tapes into core.

6-13

rT. «

1 _.

. ,

yj *

j*^ Using ^s~

s<r Extended ^n^

^s^ Memory jf

N. ? ^^

[No

Yes

Set
DF« Desired Field
IF 'Field Of BIN

Set SR> 7777

Depress LOAD ADD

High-Speed Reader

</^ Which ^S.

^ Reader ^

Low -Speed Reader

Set SR-3777

Turn TTY To LINE

Put Tope In HSR

Put Tape In LSR

Set LSR To START

Depress START

Depress CONT 1

(Object Tape \
Is Loaded J

Figure 6-4. Loading A Binary Tape Using BIN
6-14

Symbolic Editor

The Symbolic Editor is a service program which allows" the pro-
grammer to write and prepare symbolic programs and to generate a
symbolic program tape of his programs. Editor is very flexible in that
the programmer can type his symbolic program online from the Tele-
type keyboard, thus storing it directly into core memory. Then, using
certain Editor commands, the programmer can have his program listed
(printed) on the teleprinter for visual inspection.

Editor also allows the programmer to add, correct, and delete any
portion of his symbolic program. When the programmer is satisfied that
his program is correct and ready to be assembled or compiled, Editor
can be commanded to generate a symbolic program tape of the stored
program.

The Symbolic Editor program is usually issued on punched paper
tape in binary-coded format. Therefore, it is loaded into core memory
using the BIN Loader. When in core, Editor is activated for use by set-
ting the switch register (SR) to 0200 (the starting address) and de-
pressing the LOAD ADD (load address) and then START switches.
Editor responds with a carriage return/line feed sequence on the
Teletype.

Initially, Editor is in command mode, that is, it is ready to accept
commands from the programmer; anything typed by the programmer is
interpreted as a command to Editor. Editor accepts only legal com-
mands, and if the programmer types something else, Editor ignores the
command and types a question mark (?) .

When not in command mode, Editor is in text mode, that is, all
characters typed from the keyboard or tapes read in on the tape reader
are interpreted as text to be put into the text buffer in the manner spe-
cified by a preceding Editor command. Figure 6-5 illustrates how the
programmer can transfer Editor from one mode to the other.

Type a command,
then depress
RETURN Key

Type desired

input, then

CTRL/FORM Keys

Figure 6-5. Transition Between Editor Modes
6-15

Seven of Editor's basic commands are briefly described below.
Command Meaning

A Append incoming text from the keyboard into the text

buffer immediately following the text currently stored

in the buffer.
R Read incoming text from the tape reader and append

it to the text currently stored in the buffer.
L List entire text buffer; the programmer can specify one

line or a group of lines.
C Change a line; the programmer precedes the command

with the decimal line number or line numbers of the

lines to be changed.
I Insert into text buffer; the programmer specifies the

decimal line number in his program where the in-
serted text is to begin.
D Delete from text buffer; the programmer specifies the

line or group of lines to be deleted.
P Punch text buffer; the programmer can specify one

line, a group of lines, or the entire text buffer.

All commands are executed when the RETURN key is depressed ex-
cept the P command. To execute the P command, press the RETURN
key on the Teletype, turn on the punch, and press the CONT (continue)
switch on the computer console.

The above commands are only the seven basic commands. A sum-
mary of all commands is provided in Table 6-4 at the end of this section.

WRITING A PROGRAM

Now that you have some idea of what you can do with Editor and
what Editor can do for you, we will write and edit a short program, ex-
plaining each step in the comments to the right of the printout.

The example program finds the larger of two numbers and halts with
the number displayed in the accumulator (AC). The program is written
in PAL III, to be assembled using the PAL III Assembler described
later in this chapter.

The programmer loads Editor using the BIN loader (see Figure 6-4).
Editor is then activated by loading the starting address (0200 8 ) and
depressing the LOAD ADD and START switches. After Editor re-
sponds with a carriage return/line feed, the programmer types A and
RETURN key. Editor is now in text mode, that is, subsequent charac-
ters typed are appended to the text buffer. The programmer now types
the symbolic program. (Block indenting is facilitated using the CTRL/
TAB key, which Editor has programmed to indent in ten-character
increments.)

6-16

*200

CLA

/CLEAR AC

TAD NUMB

/GETB

CMA

/1'sCOMPB

IAC

/-B

TAD NUMA

/ADD -B + A

SMA.

/IF -B LARGER

. JMP .+4

/JUMP 4 LOCATIONS

CLA

/CLEAR AC

TAD NUMB

/GETB

HLT

/BIS LARGER

CLA

/CLEAR AC

TAD NUMA

/GET A

HLT

/A IS LARGER

NUMB,

0000

NUMA,

0000

Visual inspection reveals that we have errors in lines 4, 16, and 17.
(Editor maintains a line number count in decimal, with the first "line
typed being 1 and our last line being 18.) Line 4 can be removed using
the D (Delete) command, and lines 16 and 17 can be corrected using
the C (Change) command. However, Editor is presently in text mode,
and in order to issue another command Editor must be transferred to
command mode. This is done when the programmer types CTRL/
FORM (depress and hold down the CTRL key while typing the FORM
key). I

CTRL/FORM (nonprinting)

15, 16C

NUMA,
NUMB,

1111
0011

The programmer types CTRL/ FORM;
Editor responds with CR/LF and
rings the" teleprinter bell, indicating
that it is in command mode.

The programmer types 4D and the RE-
TURN key; Editor responds with a
CR/LF and the line is deleted.

The programmer types 15, 16C and the
RETURN key, informing Editor
that lines 15 and 16 (formerly 16
and 17) are to be changed.

Editor responds with a CR/LF, trans-
fers to text mode, and waits for the
programmer to change the lines.

The programmer types NUMA, 1111
and NUMB, 0011.

6-17

The symbolic program should now be correct. However, it is good
programming practice to check the program after editing; this can be
done using the L (List) command, but since only original lines 4, 16,
and 17 were changed it is not necessary to have the whole program
listed. The programmer can command Editor to list lines 4 through 17.

CTRL/FORM (nonprinting)

4, 17L

The programmer types CTRL/ FORM
to return Editor to command mode;
Editor responds with CR/LF and
rings the bell, and waits for the next
command.

The programmer types 4, 17L and the
RETURN key; Editor types lines 4
through 17.

CMA

/1'sCOMPB

IAC

/-B

TAD NUMA

/ADD -B + A

SMA .

/IF -B LARGER

JMP .+4

/JUMP 4 LOCATIONS

CLA

/CLEAR AC

TAD NUMB

/GET B

HLT

/B IS LARGER

CLA

/CLEAR AC

TAD NUMA

/GET A

HLT

/A IS LARGER

NUMA, 1111

NUMB, 0011

The changes were accepted properly. The symbolic program is correct
and ready to be punched on paper tape.

GENERATING A PROGRAM TAPE

Before issuing the P (Punch) command, Editor must be in command
mode. Figure 6-6 illustrates the procedures required to generate a sym-
bolic program tape using Editor.

The programmer types CTRL/FORM;
Editor responds with a question
mark, indicating that Editor was in
command mode.

The programmer commands Editor to
punch the entire text buffer by typ-
ing P and the RETURN key.

CTRL/ FORM (nonprinting)
?

6-18

Editor Es In

Command Mode

And

Completed Symbolic

Program Is in

Tent Buffer

Low-S

High-Speed Punc

^/^ Which >.

-<T Punch >

peed Punch

Yet
1

(

Select Switch
Register Option

j See Table 6-2

Depress HSPON

Type T And

RETURN Keys And

Depress LSP ON

Type T And
RETURN Keys

After Leader Tope
Depress LSP OFF

Type Command
IP nP Or m.nPl
find RETURN Key

•

Type Command
(PnP Or m.nPl
And RETURN Key

Depress CONT

' '

Depress LSP ON

:•

Text Is Punched

*r Punch ^S,
More Text

•S*. ? ^

Jnq

Depress CONT

. Yes

Te«f Is Punched
And Typed

Depress LSP OFF

Type FAnd
RETURN Key*

Type F And
RETURN Keys And
Depress LSP ON

Type T And
RETURN Keys

jr Punch ^>w
<£ More Text J*

Remove Tape

)

Type T And
RETURN Keys And
Depress LSP ON

Depress LSP OFF 1

After Trailer
Depress LSP OFF

Remov

I Tap«

Figure 6-6. Generating a Symbolic Tape Using Editor

When Editor recognizes a P command it waits for the programmer to
specify the low- or high-speed punch. If the programmer wants the
program punched and typed, he sets SR bit 10 to and the program
will be punched on the low-speed punch and simultaneously typed on
the teleprinter. If the programmer wants only a program tape and if he

6-19

has a high-speed punch available, he sets SR bit 10 to 1 and the pro-
gram will be punched on the high-speed punch. For the purposes of this
discussion, a printed program listing is desired, so the low-speed punch
is specified. The programmer turns on the low-speed punch and de-
presses the CONT switch on the computer console, and Editor begins
punching and typing the contents of the entire text buffer.

An image of the stored symbolic program has been punched and
typed by Editor.

If the programmer stops the computer, e.g., purposely or accidentally
turning the computer off, he may restart Editor at location 0200 or
0177 without disturbing the text in the buffer. Editor can also be re-
started at location 0176; however, all text currently in the buffer is
wiped out. Therefore, the programmer can restart at location 0176 to
re-initialize for a new program.

SEARCH FEATURE

A very convenient feature available with Editor is the search feature,
which allows the programmer to search a line of text for a specified
character. When the programmer types a line number followed by S,
Editor waits for the user to type in the character for which it is to
search. The search character is not echoed (printed on the teleprinter).
When Editor locates and types the search character typing stops, and
Editor waits for the programmer to either type new text and terminate
the line with a RETURN key or to use one of the following special
keys.

1 . «_ to delete the entire line to the left,

2. RETURN to delete the entire line to the right,

3. RUBOUT to delete from right to left one character for each

RUBOUT typed (a \ is echoed for each RUB-
OUT typed),

4. LINE FEED to insert a carriage return/line feed (CR/LF)

thus dividing the line into two,

5. CTRL/FORM to search for the next occurrence of the search

character, and/ or

6. CTRL/BELL to change the search character to the next char-

acter typed by the programmer.

INPUT/OUTPUT CONTROL

Switch register options are used with input and output commands to
control the reading and punching of paper tape. The options available
to the programmer are shown in Table 6-2. These options are used in
conjunction with the "Select Switch Register Option" operation in
Figure 6-6.

6-20

Table 6-2. Input/ Output Control

SRBit

Position

Function

Input text as is

Convert all occurrences of 2 or more
spaces to a tab

Output each tab as 8 spaces

Tab is punched as tab/rubout

Output as specified

Suppress output*

Low-speed punch and Teleprinter

High-speed punch ,

Low-speed reader

High-speed reader

♦Bit 2 allows the user to interrupt any output command and return imme-
diately to command mode; when desired, merely set bit 2 to 1.

ERROR DETECTION

Editor checks all commands for nonexistent information and incor-
rect formatting. When an error is detected, Editor types a question
mark (?), and ignores the command. However, if an argument is pro-
vided for a command that doesn't require one, the argument is ignored
and the command is executed properly.

Editor does not recognize extraneous and illegal control characters;
therefore, a tape containing these characters can be cleaned up or cor-
rected by merely reading the tape into Editor and punching out a new
tape.

SUMMARY OF SPECIAL KEYS AND COMMANDS

Using special keyboard keys and commands, the programmer con-
trols Editor's operation. Certain keys have special meaning to Editor, of
which some can be used in either command or text mode. The mode
of operation determines the function of each key. The special keys and
their function are shown in Table 6-3.

6-21

Table 6-3. Special Keys

Key

Command Mode

Text Mode

RETURN

RUBOUT

CTRL/ FORM

(period)

LINE FEED

ALTMODE
>
<

CTRL/ TAB

Execute preceding com-
mand

Cancel preceding com-
mand (Editor responds
with a ? followed by a
carriage return and line
feed)

same as *-

Respond with question
mark and remain in
command mode

Value equal to decimal

value of current line

(used alone or with +

or — and a number, e.g.

.+8)

Value equal to number of
last line in buffer; used
as an argument

List next line

List previous line

Used with . or / to ob-
tain their value

Same as = (gives value
of legitimate argument)

Enter line in text buffer

Cancel line to the left mar-
gin

Delete to the left one char-
acter for each depres-
sion; a \ (backslash) is
echoed (not used in
Read (R) command)

Return to command mode
and ring teleprinter bell

Legal text character

Used in Search (S) com-
mand to insert CR/LF
into line

Produces a tab which on
output, is interpreted as
10 spaces or a tab/rub-
out, depending on SR
option

Editor commands are given when in command mode. There are three
basic types of commands: Input, Editing, and Output. Table 6-4 con-
tains a summary of Editor commands and their function.

6-22

Table 6-4. Summary of Commands

~ Type

Command

Function

Input

Append incoming text from keyboard into text
buffer

Append incoming text from tape reader into text
buffer

Editing

List entire text buffer

List line n

m,nL

List lines m through n inclusively

Change line n

m,nC

Change lines m through n inclusively

Insert before first line

Insert before line n

Delete entire text buffer

Delete line n

m,nD

Delete lines m through n inclusively

m,n$kM

Move lines m through n to before line k

Print next tagged line (if none, Editor types ?)

Print next tagged line after line n (if none, ?)

Search buffer for character specified after RE-
TURN key and allow modification (search char-
acter is not echoed on printer)

Search line n, as above

m,nS

Search lines m through n inclusively, as above

Output

Punch entire text buffer

Punch line n

m,nP

Punch lines m through n inclusively

Punch about 6 inches of leader/ trailer tape

Punch a FORM FEED onto tape

Do P, F, K, and R commands

m and n are decimal numbers, and m is smaller than n; K is a decimal number.

The P and N commands halt the Editor to allow the programmer to select
I/O control; press CONT to execute these commands.

Commands are executed when the RETURN key is depressed, excluding the
P and N commands.

Symbolic Assemblers

A symbolic assembler is a service program that translates symbolic
programs into binary-coded programs which can be loaded and run
on the computer. In other words, the programmer writes the symbolic
program using symbols which are meaningful to him, and then an as-
sembler is used to translate the symbols into binary code, which is
meaningful to the computer. The computer knows only yes or no, plus

6-23

or minus, voltage or no voltage, magnetized or demagnetized, i.e., one
of two conditions (states) which we simplify as 1 or 0. Therefore, the .
assembler translates the programmer's symbols into l's and 0's which
are meaningful to the computer.

Because assemblers are vital to the efficient operation of computers
in the PDP-8 family, DEC presently offers four (PAL III, MACRO-8,
PAL-D, and 8K SABR), and this list is destined to grow with time as
other versions are needed to assemble other programming languages.
This section includes general descriptions of PAL III and MACRO-8.
PAL-D incorporates most of the features of both PAL III and
MACRO-8 and is used only in the Disk Monitor System. So, if you
learn PAL III and MACRO-8, you have only to learn the few excep-
tions of PAL-D and then you know all three.

First PAL III is discussed and then MACRO-8. MACRO-8 is com-
patible with PAL III except for some additional features, therefore, in
the section on MACRO-8 emphasis is on the additional features and
exceptions. .

PAL III SYMBOLIC ASSEMBLER

The PAL III Symbolic Assembler (PAL stands for Program Assembly
Language) is an indispensable service program used to translate symbolic
programs,, which are written in the PAL III language, into binary-coded
programs (binary programs). Having progressed to this section of the
Handbook, you are by now familiar with the PAL III programming
language; because the symbolic language used in the preceding chapters
is PAL HI. In this section, the PAL III Assembler is used to assemble
the example program written using Editor (see "Writing a Program,"
above).

PAL III is a two-pass assembler with an optional third pass, i.e., the
symbolic program tape must be passed through the assembler two times
to produce the binary-coded tape (binary tape), and the optional third
pass produces a complete octal/ symbolic program listing which can be
typed and/or punched if desired. A brief explanation of the three passes

is given below.

Pass 1. The assembler reads the symbolic program tape and defines all

symbols used and places these user symbols in a symbol table for use
during Pass 2. The assembler checks for undefined symbols and certain
other errors and types an error message on the teleprinter when an error
is detected.

Pass 2. The assembler rereads the symbolic program tape and gener-
ates the binary tape using the symbols defined during Pass 1. When the
low-speed punch is used, meaningless characters will be typed on the

6-24

teleprinter, and these should be ignored by the programmer. The assem-
bler checks illegal referencing during this pass andtypes an error mes-
sage on the teleprinter when any is detected.

Pass 3. The assembler reads the symbolic program tape and types
and/or punches the octal/symbolic program assembly listing. This list-
ing thoroughly documents the assembled program and is useful when
debugging and modifying the program.

The meaningless characters, error messages, and octal/symbolic pro-
gram listing will be shown later in this section.

PAL III accepts symbolic program tapes from either the low-speed
or high-speed reader and produces the binary tapes on either, the low-
speed or high-speed punch.

During assembly, the programmer communicates with PAL III via
the switches on the computer console. Switch options are used to specify
which pass the assembler is to perform and which reader and punch the
assembler should accept input from and punch out on.
ASSEMBLING A SYMBOLIC PROGRAM. Earlier in this chapter,
the programmer wrote a PAL III symbolic program and generated the
symbolic program tape using Editor. That symbolic program can now
be assembled to produce a binary program using PAL III. A listing of
the symbolic program follows.

*200

CLA /CLEAR AC

TAD NUMB /GETB

CMA /l'SCOMPB

IAC /— B

TAD NUMA /ADD -B + A

SMA /IF -B LARGER

JMP .+4 /JUMP 4 LOCATIONS

CLA /CLEAR AC

TAD NUMB /GET B

HLT /B IS LARGER

CLA /CLEAR AC

TAD NUMA /GET A

HLT /A IS LARGER
NUMA, 1111
NUMB, 0011

First, PAL HI must be loaded into core memory, and since PAL III
is on punched paper tape in binary-coded format, it is loaded into core
memory using the BIN Loader (see Figure 6-4 for loading procedures).

6-25

With PAL III in core, we are ready to assemble the^ symbolic pro-
gram. Figures 6-7 and 6-8 illustrate the procedures for assembling with
PAL III using the low-speed reader/punch and high-speed reader/
punch, respectively. In these flowcharts, the switch register options are
set for the appropriate reader/punch.

The low-speed reader and punch (LSR and LSP) are used in the
following assembly (see Figure 6-7).

Initializing and Starting

Entering Pass 1

NUMA
NUMB

0215
0216

Entering Pass 2
BB:

8 8

<:
<)
Entering Pass 3

Load PAL III into core memory using

BIN.
Set SR=0200 and depress LOAD ADD.
Turn TTY to LINE and put symbolic

program tape in LSR.
Set SR=2200 and set LSR to START.
Depress LSP to ON and depress

START.
Error messages would be typed now.
Symbol table concludes Pass 1 .
Put symbolic program tape in LSR.
Set SR=4200 and set LSR to START.
Depress LSP to ON and depress CONT.
Disregard meaningless characters while

object tape is being punched.
Error messages would be typed now.
Put symbolic program tape in LSR.
Set SR=6200 and set LSR to START.
Depress LSP to ON and depress CONT.
The octal/ symbolic program listing is

being typed and punched.

*200

0200

7200

CLA

/CLEAR AC

0201

1216

TAD NUMB

/GETB

0202

7040

CMA

/l'SCOMPB

0203

7001

IAC

/-B

0204

1215

TAD NUMA

/ADD-B + A

0205

7500

SMA

/IF -B LARGER

0206

5212

JMP .+4

/IUMP 4 LOCATIONS

0207

7200

CLA

/CLEAR AC

0210

1216

TAD NUMB

/GETB

0211

7402

HLT

/B IS LARGER

0212

7200

CLA

/CLEAR AC

0213

1215

TAD NUMA

/GET A

0214

7402

HLT

/A IS LARGER

0215

1111 NUMA, 1111

0216

0011 NUMB, 0011

NUMA

0215

NUMB

0216

6-26

(D-A

D«pr«ii LOAD ADO

Turn TTY To LINE

r" : " 1 _ ■ Pot Syr

| O.pr... STOP M -^To^ln

Put Symbolic
n LSR

Figure 6-7. Assembling with PAL III Using Low-Speed Reader/Punch

6-27

Set SR*2201

Punch ' X^ Ye*
Symbol Toble

Went For

Symbol Table

On TTY

O-t

Oeprew HSP

POWER To OH

Am) FEED

-\ See Figure 6-4

Set

OF "Desired Field
IF. Field 01 PAL m

Decrees LOAD ADD

Turn TTY To LINE

Depress STOP

Depress MSP

POWER To ON

And FEED

Wail For End
Of Binary Tape

Sir SR=6200

Print ^SPunch

Depress HSP

POWER To ON

And FEED

Wait For
End Of Listing

Figure 6-8. Assembling with PAL III Using High-Speed Reader/Punch

6-28

The tape produced during Pass 2 is the binary tape, which is loaded
into core memory using the BIN Loader. The symbol table tape pro-
duced during Pass 1, the binary tape produced during Pass 2, and the
octal/symbolic program listing produced during Pass 3 are used when
debugging the program.

PSEUDO-OPERATORS. When writing a PAL III symbolic pro-
gram, the programmer can use pseudo-operators (pseudo-ops) to direct
the assembler to perform certain task or to interpret subsequent coding
in a certain manner. Some pseudo-ops generate storage words in the
binary program, others direct the assembler on how to proceed with the
assembly. Pseudo-ops are maintained in the assembler's permanent
table, 1 which can be altered using certain psuedo-ops.

When using more than one memory bank, the pseudo-op FIELD in-
structs the assembler to output a field setting:

FIELD n where n is an integer, a previously defined symbol, or a
symbolic expression within the range through 7,
inclusive.

Integers used in a symbolic program are usually taken as octal num-
bers. However, if the programmer wishes to have certain numbers
treated as decimal, he may use the pseudo-op DECIMAL, and then re-
turn to the original radix with the pseudo-op OCTAL.

DECIMAL all integers in subsequent coding are taken as decimal

until the occurrence of the pseudo-op OCTAL,
OCTAL which resets the radix to its original base.

In an indirect address instruction, the special symbol "I" between the
operation code and the operand, or address field, is another pseudo-op.
(Indirect addressing was covered in Chapter 2.)

When a symbolic program is very long, it is often desirable to punch
the symbolic program on two or more physical lengths of tape. The last
instruction of each section must be the pseudo-op PAUSE, and the
last instruction of the program must be a dollar sign ($).

PAUSE stops the assembler, but the current pass is not ter-

minated. When the programmer is ready to assemble
the next section, he has only to depress CONT on the
computer console. (PAUSE is normally used only at
the physical end of tape. )

iThe permanent symbol table contains operation codes (MRI's, IOTs, micro-
instructions, and pseudo-ops) and their octal equivalents; it is a permanent
part of the assembler. An external symbol table contains user-assigned sym-
bols and their corresponding absolute addresses; it is created during assembly
and punched on the binary tape during Pass 2.

6-29

The last three pseudo-ops are used to alter the permanent symbol
table. They are:

EXPUNGE erases the entire permanent symbol table excluding the
pseudo-ops.

FIXMRI meaning FIX Memory Reference Instructions. The

pseudo-op FIXMRI must be followed by one space,
the symbol for the MRI to be defined, an equal sign,
and the octal value of the symbol to the -immediate
left of the equal sign.

FIXTAB meaning FIX the current permanent symbol TABIe. All

symbols that have been defined before the occurrence
of this pseudo-op are made part of the permanent
symbol table.

Therefore, with these three pseudo-ops the programmer can alter the
permanent symbol table to contain only those symbols he needs to
assemble his symbolic program, which in turn provides more core for
the external symbol table.

OUTPUT CONTROL. Output is controlled by the setting of switch
register bit 11. When the assembler first enters a pass, it "looks" at the
state of SR bit 11 and outputs as specified. The SR bit 11 options are:

Bit 11=0 Output on teleprinter and low-speed punch.
Bit 11 = 1 Output on high-speed punch.

ERROR MESSAGES. The assembler is constantly checking for assem-
bly errors, and when any is detected, an error message is printed on the
teleprinter. Error messages are printed in the following format.

xx yyyyyy AT nnnn

where xx is the error message (see below), yyyyyy is the symbol or
octal value of tbe symbol of the error occurring AT location nnnn.

Assembly errors are checked for during Pass 1 and Pass 2 only. The
error codes are listed below.

Pass 1:

Illegal Character

Redefinition

Duplicate Tag

Symbol Table Full

Undefined Address

Illegal Reference

Pass 2:

For a thorough description of PAL III, see PAL III Symbolic As-
sembler, Order No. DEC-08-ASAC-D.

6-30

MACRO-8 SYMBOLIC ASSEMBLER

The MACRO-8 Symbolic Assembler is a service program used to
translate symbolic programs that are written in the MACRO-8 symbolic
language into binary programs. The MACRO-8 language can be gen-
erally considered as PAL III with the following additional features:

1. User-Defined Macros — Groups of computer instructions re-
quired for the solution of a specific problem can be defined by
the user as a macro instruction (explained later).

2. Double Precision Integers — Positive or negative double preci-
sion integers are allotted two consecutive core locations.

3. Floating-Point Constants — The format and rules for denning
these constants are compatible with the format used by the
Floating-Point System.

4. Operators — Symbols and integers may be combined with a
number of operators.

5. Literals — Symbolic or integer literals (constants) are auto-
matically assigned.

6. Text Facility — There are text facilities for single characters and
blocks of text.

7. Link Generation — Links are automatically generated for off-
page references.

With these additional features, it is clear that the MACRO-8 Assem-
bler requires more core memory than the PAL III Assembler. There-
fore, programs originally coded to be assembled by PAL III might have
too many user symbols to be assembled by MACRO-8 (it was necessary
to decrease the size of the user's symbol table to incorporate the addi-
tional features). However, the programmer can, using the appropriate
pseudo-op, increase or decrease the size of the permanent symbol table
if desired.

MACRO-8 is a two-pass assembler with an optional third pass which
produces a octal/ symbolic program assembly listing. There are two
versions of MACRO-8.

The Low Version uses the low-speed reader for all input and the
teleprinter and low-speed punch for all output.

The High Version uses the high-speed reader for all input, the high-
speed punch for binary output (and, if desired, the program listing
tape), and the teleprinter and low-speed punch for output of error
diagnostics, symbol table, and the third pass program listing.

6-31

The three passes of MACRO-8 are identical to those of PAL III, in
fact, the example program assembled by PAL III in the preceding sec-
tion could be assembled by MACRO-8, and the printed output would
be the same for both assemblers.

MACROS. When writing a program, it often happens that certain
coding sequences are used several times with just the arguments
changed. If so, it is convenient if the entire sequence can be generated
by a single statement. To do this, the coding sequence is defined by
dummy arguments as a macro. A single statement referring to the macro
by name, along with a list of real arguments, will generate the correct
sequence in line with the rest of the coding.

The macro name must be defined before it is used. The macro is de-
fined by means of the pseudo-operator DEFINE followed by the
macro's name and a list of dummy arguments. For example,

DEFINE MOVE DUMMY1 DUMMY2

<CLA

TAD DUMMY 1

DCA DUMMY2

TAD DUMMY2>

The actual choice of symbols used as dummy arguments is arbitrary;
however, they may not be defined or referenced prior to the macro
definition. The actual definition of the macro must be enclosed in angle
brackets ( < > ).

The above definition of the macro MOVE could have been, coded as
follows:

DEFINE MOVE ARG1 ARG2

<CLA; TAD ARG1; DCA ARG2; TAD ARG2>

When a macro name is processed by the assembler, the real argu-
ments will replace the dummy arguments. For example, assuming that
the macro MOVE has been defined as above,

*400

A, 0400 0000

B, -6. 0401 7772
MOVE A, B 0402 7200
$ 0403 1200

0404 3201

0405 1201

6-32

A macro need not have any arguments. A sequence of coding to
rotate the contents of the accumulator and link six places to the left
might be stated as follows.

DEFINE ROTL6
<RTL; RTL; RTL>

A macro is referenced by giving the macro name, a space, and then
the list of real arguments, separated by commas. There must be at least
as many arguments in the macro reference as in the corresponding
macro definition. For a detailed list of other macro restrictions, see
MACROS Assembler, Order No. DEC-08-CMAA-D.
LITERALS. Since the symbolic expressions which appear in the ad-
dress part of an instruction usually refer to the address of locations con-
taining the quantities being operated upon, the programmer must ex-
plicitly reserve the locations holding his constants. The MACRO-8 pro-
gramming language provides a means for using a constant directly. Sup-
pose, for example, that the programmer has an index which is incre-
mented by two. One way of coding this operation would be as follows.

CLA

TAD INDEX
TADC2
DCA INDEX

C2, 2
Using a literal, this would become

CLA

TAD INDEX
TAD (2)
DCA INDEX

The left parenthesis is a signal to the assembler that the expression fol-
lowing is to be evaluated and assigned a location in the constants table
of the current page. This is the same table in which the indirect address
linkages are stored. In the above example, the quantity 2 is stored in a
location in a list beginning at the top of the memory page (page address
177), and the instruction in which it appears is encoded with an ad-
dress referring to that location. A literal is assigned to storage the first
time it is encountered, and subsequent references will be to the same
location.

6-33

If the programmer wishes to assign literals to page rather than the
current page, he may use square brackets, [ and ], in place of the paren-
theses. However, in both cases, the right or closing member may be
omitted, as in the example below.

Literals may be nested. For example:

*200

TAD (TAD (30

will generate

0200

1376

0376
0377

1377
0030

This type of nesting may be carried to as many levels as desired.
PSEUDO-OPS. MACRO-8 has available a number of useful pseudo-
ops, which are listed and briefly explained below.

PAGE When used without an argument, the current location

counter is reset to the first location on the next succeed-
ing page. With an argument (PAGE n), the current
location counter is reset to the first location on the spec-
ified page, page n.

DECIMAL When this pseudo-op occurs, all integers encountered in
subsequent coding will be taken as decimal until the
occurrence of OCTAL,

OCTAL which will reset the radix to its original base.

PAUSE When several tapes are to be assembled together, each

except the last (which ends with $) should have as its
last symbol the pseudo-op PAUSE. This causes the
assembler to stop processing and halt the computer.
After placing a new tape in the tape reader, assembly
can be continued by depressing CONT on the computer
console.

EXPUNGE When used, the entire permanent symbol table (exclud-
ing pseudo-ops) is erased. This pseudo-op is used when
the programmer wishes to provide more core for the
user program symbols. Then, with FIXTAB,

FIXTAB the programmer can build a customized permanent sym-

bol table, containing only those permanent symbols re-
quired for his user program.

DEFINE Used to define macros. This pseudo-op was covered un-
der "Macros," above.

The preceding pseudo-ops and a few others are described in detail in
the MACROS Assembler manual.

6-34

OFF-PAGE REFERENCING. During assembly, the page bits of the
address field are compared with the page bits of the current location
counter (see "Page Addressing" in Chapter 2). If the page bits of the
address field are nonzero and do not equal the page bits of the current
location counter, an off -page reference is being attempted (see "Indirect
Addressing" in Chapter 2). If the reference is to an address not on the
page where the instruction will be located, the assembler will set the
indirect bit (bit 3) and an indirect address linkage will be automatically
generated on the current memory page.-

Although the assembler will recognize and automatically generate an
indirect address linkage when necessary, the programmer may still indi-
cate an explicit indirect address using the special symbol "I" between
the operation code and the address field.

As can be seen by comparing Figures 6-9 and 6-10 to Figures 6-7
and 6-8, the assembly procedures for MACRO-8 and PAL III are
similar but certainly not identical.

SWITCH REGISTER OPTIONS. As the assembler begins to enter
each pass, it "reads" or "looks at" or "senses" (take your choice) cer-
tain bits of the switch register which tell it how to proceed. The optional
settings of the switch register are shown in the table below.

Table 6-5. Switch Register Options

SRBit

Position

Meaning

0-11

Enter next pass.

Erase symbol table, excluding permanent symbol, and
enter Pass 1 (the programmer then depresses STOP
andCONT).

Enter Pass 2 to generate another binary tape.

. 2

Enter Pass 1 without erasing denned symbols.

Enter Pass 3 .

Delete double precision integer and double precision
floating-point processors (this increases the symbol
table size by 100 8 symbols).

Delete macro and number processors (this increases
the symbol table size by 175 8 symbols).

Switch register bits 10 and 11 are sensed by MACRO-8 when it enters
Pass 1, and remembers these settings throughout the assembly. If these
bits are set to 1, MACRO-8 would have to be reloaded to handle sub-
sequent programs that use macros, double precision integers, or float-
ing-point numbers.

6-35

0-{

Load MACRO-8

I 1 Stt Figure 6-'

DF-Owirtd Field
IF ■ FWdOt MACROS

I PulTmnInLSR 1 X

^ | Put Tope In LSR | jj

J 4 | Put LSR To START |

©H

— <D

Figure 6-9. Assembling a MACRO-8 Symbolic Program
Using the Low-Speed Reader/Punch

6-36

(•>-[

Lood MACRO-8

(£H*C

1 C

S« Fljgf. 6-4

Srt SR-O200

Set SR-4200

Put Tope In MSR

D*pr*ss HSP
POWER To ON

I P«pf«M HSP FEED [
"•j P«pr«M START |*~

_£

Depress CONT

I P»l Taptl,, HSR I I Tap. Stop, |

Depress STOP

©*-c

D«pres*STOP

Set Location
0004 « 0600

Dtpttst HSP
POWER To OFF

Dtprm t-SP ON I

Correct Error
Uting Editor

-©

Figure 6-10. Assembling a MACRO-8 Symbolic Program
Using the High-Speed Reader/Punch

6-37

When assembling using the high-speed reader, the Pass 3 program
listing is normally output on the teleprinter and low-speed punch. How-
ever, when assembling a long program, the programmer may wish to
have the third pass listing output on the high-speed punch. This can be
done by changing the contents of location 0004 from 2600 to 0600
(see Figure 6-10). It is advised that this change not be made until Pass
3, so that Pass 1 and 2 diagnostics will be printed.

ERROR MESSAGES. The assembler is constantly checking for as-
sembly errors, and when one is detected, an error message is printed on
the teleprinter. Error messages are printed in the following format.

xx yyyyyy

where xx is a two-letter code which specifies one of the types of errors
listed below, and yyyyyy is either the absolute octal address where the
error occurred or the address of the error relative to the last symbolic
tag (if there was one) on the current page.

Following is a descriptive list of the error messages, some are printed
out during Pass 1 , others during Pass 2, and some of which are printed
out during both passes.

Error Code Meaning

BE MACRO-8 internal tables have overlapped

IC Illegal character

ID Illegal redefinition of a symbol

IE Illegal equal sign

II Illegal indirect address

IM Illegal format in a macro definition

LG Link generated to off-page address*

MP Missing parameter in macro call

PE Current, nonzero page exceeded

SE Symbol table exceeded

US Undefined symbol

ZE Page zero exceeded

*This is to inform the user of off-page references which may not be an error.
This diagnostic can be suppressed to speed up Pass 2 assembly by setting
location 1234 to 7200 prior to entering Pass 1.

For a thorough description of MACRO-8, see MACROS Assembler,
Order No. DEC-08-CMAA-D.

6-38

CONCLUSION

The programmer decides which assembler is needed to translate his
symbolic program into a binary program. The decision depends on such
things as the amount of core memory available, the size of the symbolic
program and the number of user symbols it contains, and whether the
symbolic program contains macros, literals, etc.

Dynamic Debugging Programs

Included in the System Software package are two dynamic debugging
programs: DDT-8 (Dynamic Debugging Technique) and ODT-8 (Oc-
tal Debugging Technique). Dynamic debugging programs are service
programs which allow the programmer to run his binary program on
the computer and use the Teletype keyboard to control program execu-
tion, examine registers, change their contents, make alterations to the
program, and much much more.

A symbolic program can be assembled correctly and still contain
logical errors, i.e., errors which cause the program to do something
other than what is intended. The assembler checks for certain syntax
errors, not logical errors. Syntax errors include undefined tags, mis-
spelled tags and operation codes, and incorrect format (e.g., omission
of a required operand). Logical errors are detected only when the pro-
gram is running on the computer.

DEBUGGING WITHOUT DDT-8 OR ODT-8

If the programmer feels sure that his program is correct and ready
for use, he can simply load the program and let it run until it stops (if it
stops). And if the program doesn't produce the correct results, the pro-
grammer without a DDT program can use the console switches to
examine- specific locations one-by-one to try to find the error(s) by
interpreting the console lights. There are two hazards to this approach.
First, by the time the program stops the error may have caused all
pertinent information, including itself, to be altered or eliminated.
Second, the program may not stop at all, it might continue to run in
an infinite loop; such loops are not always easy to detect.

Added to these problems are the difficulties of interpreting binary
console displays and translating them into symbolic expressions related
to the user's program listing. Further, adding corrections to a program
in the form of patches (altered and added instructions or routines) re-
quires seemingly endless manipulation of the console switches. In all
this, the chance of programmer error at the console is large and is likely
to obscure any real gain made from debugging.

6-39

The programmer can, of course, while sitting at his desk using the
program assembly listing, mentally execute his program. This method
is frequently used with very short programs, very short only — human
memory can not retain every step and instruction in even a fairly short
program; it can not match computer memory.

What is needed to conveniently and accurately debug a user program
is a service program which will assume the tasks the programmer would
have to perform if he used the console switches. DDT-8 and ODT-8
are such debugging programs.

DEBUGGING WITH DDT-8

Using DDT-8, the programmer can run his program on the computer,
control its execution, and make corrections to the program by typing
commands to DDT-8 and altering his program from the Teletype key-
board. Tracking down a subtle error in a complex section of coding is
a laborious and frustrating job if done by hand, but with the breakpoint
facility (explained later) of DDT-8, the user can interrupt the operation
of his program at any point and examine the state of the program and
computer. In this way, sources of trouble can be isolated and quickly
corrected.

By the time the programmer is ready to start debugging a new pro-
gram at the computer, he should have at the console:

1 . The binary tape of the new program.

2. The symbol definition tape which was part of the assembly
output from Pass 1.

3. A list of the symbols and their definitions.

4. A complete octal/symbolic program listing.

5. A binary tape of the DDT-8 program, which is loaded into
core memory using BIN (see Figure 6-4) .

To begin the debugging run, first ascertain that BIN is in memory,
then load the programmer's binary program and the binary DDT-8 pro-
gram tape (see Figure 6-4 for loading procedures). Figure 6-11 illus-
trates the procedures for loading and executing DDT-8.

Core memory now contains the DDT-8 program, the user's program,
and a table of permanent symbol definitions. This table includes the
definitions for all of the memory reference instructions, microinstruc-
tions, the ten basic IOT instructions, and the combined operations CIA
and LAS.

6-40

(

h -

Looa
Object Program ,

-| See Figure 6-4

(See Figure 6-12)

—

Load DDT-8

-J See Figure 6-4

Set SR«5400

Depress LOAD ADO

Depress START

)

(

)ebug Object Program 1
DDT-8 Is In
Command Mode ,

— | See

Figure

6-12

Figure 6-11. Loading and Executing DDT-8

The programmer may communicate with his program using either
the mnemonic coding of the symbolic program or the octal coding of
the binary program. Therefore, since DDT-8 is to perform all transla-
tion between binary and symbolic representation, it must have access
in memory to the user's symbol definitions. The external symbol table
tape (containing the user's symbol definitions) produced during as-
sembly at the end of Pass 1 is now loaded into memory by DDT-8.
Figure 6-12 illustrates the steps required in loading the Pass 1 symbol
table tape.

The external symbols of the user's program are stored in memory im-
mediately below DDT-8's permanent symbol table (mentioned in the
paragraph below Figure 6-11). These symbols, and any others which
may be entered from the console constitute the external symbol table.
Additional symbols may be appended to the external symbol table by
performing the steps illustrated in Figure 6-13.

A new external symbol tape can be generated off-line using the Tele-
type keyboard and low-speed punch merely as a tape preparation fa-
cility. The symbol tape can then be loaded into memory using DDT-8
as shown in Figure 6-12.

6-41

J—

•\ See Figure 6 -11

Turn TTY To LINE

Put Symbol Table
Tope In LSR

Type t R

Set LSR To START

(■>

(See Figure 6-11)

^_ Reads In J>
TVes

Tope Stops

Set LSR To FREE

Depress CONT

Octal Address Typed

Is Lower Limit Of

External Symbol Table

^r More ^^*n.

^s. ? ^/^

Jno

Yes

r Externa
\_ Table I

Symbol ^
i Loaded J

Figure 6-12. Loading the External Symbol Table Tape Using DDT-8

After loading the external symbol table tape, the address typed out
will be the lowest memory location which is occupied by a symbol
definition. The Programmer is now ready to begin debugging using
DDT-8.

The example program to be checked out is a subroutine which ac-
cumulates the sum of the first n integers. The program is shown below.

6-42

(

^ With DDT-8 "\
^ In Command Mode J

1 ■

Load External
Symbol Table

Set SR-1400

Type CR

Type CR/LF

Type New Symbol

Type 1 Or More Spaces

Type Octal Value

Of Definition Of

New Symbol

Yea

^^ More ^n.
^^ Symbols /^

Tno

Type CR/LF And EOT

Depress CONT

f Address Typed ^

Is Lower Limit Of ]

\External Symbol Table/

Figure 6-13. Appending New Symbols to External Symbol Table

6-43

/INTEGER SUMMATION SUBROUTINE

INTSUM,

CLA

TAD I INTSUM

/GET COUNTER

DCAN

/SAVE IT

DCA PSUM

/CLEAR PSUM

LOOP,

TAD PSUM

/GET SUM

TAD INT

/ADD INTEGER

DCA PSUM

/SAVE NEW SUM

ISZN

/DECREMENT COUNTER, N=0?

JMP LOOP

/NO, NOT FINISHED

TAD PSUM

/YES, FINISHED. AC^SUM

ISZ INTSUM

/GET OVER ARG.

IEXIT,

JMP I INTSUM

/RETURN

PSUM,

*300

ITEST,

CLA

JMS INTSUM

/INTSUM TEST PROGRAM

INT,

/PUT INTEGER HERE

RTN,

HLT

For testing purposes, a short calling sequence has been included
which provides the integer limit of the sum as an item of data. The first
task is to place an integer in the core location that holds this datum,
namely, the location labeled INT in the calling program. By typing the
address of the core location (which in this case can be done by typing
the address tag), followed by a slash, the programmer indicates to
DDT-8 that he wishes to examine the contents of that location. Thus,
he types

INT/

and DDT-8 responds to the slash by typing an expression which has
the value of the contents of the specified location. In this case, the con-
tents of INT = 0, and the line now appears as follows.

INT/ AND 0000

NOTE: In the examples, information typed by DDT-8 is underlined to in-
dicate a distinction from information typed by the programmer. In
actual operation no underlining is present.

After typing the contents of the specified core location, DDT-8 types
five spaces and waits. The location is now open, which means that its
contents are available for modification. The programmer decides that
the first test integer is 10. This must be an octal integer since DDT-8

6-44

performs no decimal arithmetic. With the location open, the program-
mer types the number 10. Then, to close the location, he types a car-
riage return (RETURN key) immediately after the number.

INT/ AND 0000 10 (RETURN key was typed after 10)

Further access to this location is now denied until the programmer
opens it again.

Having provided his data, the programmer is ready to start the pro-
gram. If it works, it should stop almost immediately with the sum of
the first 10 8 integers, which is 100 8 , displayed in the accumulator (AC)
lights. To start the program, the programmer types the following com-
mand.

ITEST[G

The left bracket ([) is printed when the ALT MODE key is typed; its
function here is to identify the succeeding character as a DDT-8 com-
mand. The letter G specifies the action to be performed, which in this
case causes DDT-8 to transfer control to the test program at location
ITEST.

The programmer has typed the G command; his program starts
executing. Immediately he observes that something is wrong — the pro-
gram did not stop almost instantaneously, but ran for a very short, but
observable, time. The contents of the AC lights are definitely not equal
to 100 8 .

The programmer restarts DDT-8 after each program execution, that
is, he sets the SR to 5400 and then presses LOAD ADD and START
switches. He must also open the PSUM and N Locations to restore
them to zero before continuing debugging.

At this point, the programmer knows something is wrong, but he is
not sure where the error lies. If he could interrupt the program during
its operation, he might get a better idea of the nature of the problem.
For instance, if he could verify that the data was transferred to the
subroutine correctly, he could eliminate the calling sequence as a source
of error.

The DDT-8 facility which allows the programmer to interrupt the
operation of the program at any time is called the breakpoint. As its
name implies, it allows him to break into the program sequence at some
point and return control to DDT-8. He can specify a breakpoint by
typing the address of the instruction where he wants to interrupt the
sequence, and after this -address he types the breakpoint command
( [B ) . If he requests a breakpoint at location

6-45

INTSUM+3

the program will be interrupted when the datum is in the AC, but be-
fore it is deposited in the working location n. The breakpoint command
would follow and the statement would appear as follows.

INTSUM+3 [B

When this command is given, the information is retained by DDT-8
until the programmer executes the program. At that time, the sequence
of operations performed my DDT-8 is:

1.- The contents of location INTSUM+3 are saved in a special
location within DDT-8.

2. In place of the instruction in location INTSUM+3, DDT-8
substitutes the instruction JMP I 4. Location 4 contains the
address of a special breakpoint handling subroutine within
DDT-8.

3. After the breakpoint has been set up, DDT-8 passes control
to the programmer's program.

To ascertain that the error did not destroy the item of data in the
calling program, check it by opening the location.

INT/ AND 0010

The programmer then opens the PSUM and N Locations and restores
them to zero.

Having ascertained that the datum is correct, start the program again.

1TEST[G

Almost immediately, the breakpoint is encountered and control re-
turns to DDT-8. When the breakpoint occurs, DDT-8 saves the con-
tents of the AC. It then types the address of the breakpoint, a right
parenthesis, and the contents of the AC which have been saved.

INTSUM+0003)0010

The programmer sees that the transfer is correct. Therefore he restarts
DDT-8 to continue debugging.

In similar fashion, the programmer moves the breakpoint to the end
of the subroutine at location IEXIT. He discovers that at this point
the error has manifested itself. He knows now that the trouble is in the
initialization of the main loop. The breakpoint is set at LOOP to dis-
cover that the datum is placed in location n as desired. Now the break-
point is moved to the end of the loop.

6-46

L00P+3[B

ITEST[G

LOOP+0003)D010

At the end of the first pass through the loop, the contents of the AC
are equal to the starting value of n. At this point, however, the contents
of n itself have just been changed. If the subroutine is working properly,
the contents of n should now be equal to 7770. The programmer in-
vestigates:

N/ AND 0010

The programmer now realizes what he did wrong. In his attempt to
save space by using the datum as a counting index, he forgot that the
ISZ instruction increments the contents of negative value. Therefore,
he must negate the counter by inserting the CIA instruction as shown
below.

INTSUM,

CLA

TAD I INTSUM

/GET COUNTER

CIA

/NEGATE COUNTER

DCAN

/SAVE IT

and the program will run properly. Realizing this, the programmer con-
cludes the debugging session by typing

to remove the breakpoint. When the breakpoint is removed, the original
contents of the break location is restored.

The example debugging session above was. simple; the error was
obvious. This is seldom the case, and with long or complex programs
several debugging runs may be required. Being able to debug a program
using symbolic expressions shortens the time required to arrive at a
correct, workable program.

DEBUGGING NOTES. For DDT-8 to be useful, the programmer
should be aware of and consider certain factors. Namely, the following:

6-47

1. Do not open any symbol table location.

2. The symbol table tape is loaded using the low-speed reader
only.

3. Each user symbol occupies four locations in the symbol table
storage area.

4. Input is interrupted when the symbol table storage area is full.

5. To enter a combined operate class IOT instruction into an open
location, the combination must contain no more than two
mnemonics, the second of which must be CLA. Any other
combination is illegal and therefore ignored.

6. DDT-8 is restarted at its starting address, 5400, unless the
programmer wishes to restart before he has punched a com-
plete tape with checksum. In which case, he must restart at
location 5401, and the checksum is preserved.

DDT-8 offers a very convenient feature, the word search feature,
which allows the programmer to look at the contents of one, a group,
or all locations in his program. A word search is indicated to DDT-8
when the programmer types the upper and lower limits of the search
using the [L and [U commands and then the word search command,
nnnn[W (nnnn is the word being searched). Address and location con-
tents are printed as symbolic expressions or octal integers, according
to the mode (symbolic or octal) specified by the programmer using
the [S or [O command. The search never alters the contents of any
location examined. The word search feature is detailed in DDT-8,
Order No. DEC-08-CDDA-D.

ERROR DETECTION. DDT-8 is constantly watching for certain
errors, those listed below.

1 . Undefined symbol or illegal symbol

2. Illegal character

3. Undefined control command

4. Off-page addressing

When an error is detected, DDT-8 types a question mark (?) on the
teleprinter and ignores all the information typed between the point of
the error and the previous tab or carriage return.
SUMMARY OF SPECIAL KEYS AND COMMANDS. The program-
mer controls DDT-8 from the keyboard using special keys and com-
mands. Certain keys have special meaning to DDT-8, and they must

6-48

be typed in certain formats in order to be acceptable to DDT-8. Tables
6-6 and 6-7 summarize the special keys and commands.

Table 6-6. DDT-8 Special Keys

Key

Meaning

(space)

Separation character

+ (plus)

Specifies address arguments relative to symbols

— (minus)

Same as +

(period)

Current location; used in address arguments

= (equal)

Type last quantity as an octal integer

RETURN

Make modifications, if any, and close location

LINE FEED

Make modifications, if any, close location and open

N '

next sequential location

/ (slash)

Location examination character; when following the

address location, the location is opened and its con-

tents printed

t (up-arrow)

When following a location printout, the location

addressed therein is opened

<- (back-arrow)

Delete the line currently being typed

Table 6-7. DDT-8 Commands

Command

Meaning

Mode Control

[S
Input

Set DDT-8 to type out in octal
Set DDT-8 to type out in symbolic

Read symbol tape into external table from LSR, or
define new symbol from keyboard
Program Examination and Modification

nnnn/

RETURN

LINE FEED

t (SHIFT/I)

Open location nnnn (nnnn may be octal or sym-
bolic)

Close location currently open; enter modification,
if any

Close location currently open and open next se-
quential location; enter modification, if any
Close location currently open and open location
address therein; enter modification, if any

6-49

Table 6-7. DDT-8 Commands (Cont.)

Command

Meaning

Breakpoint Insertion and Control

nnnn[B

nnnn[G

n[C

Word Search
nnnn[W

Remove current breakpoint

Insert a breakpoint at location nnnn

Go to location nnnn and start program execution

Continue from breakpoint, execute breakpoint n

times and return control to programmer. If n is

absent, it is assumed to be 1 .

Begin word search for all occurrences of expres-
sion nnnn masked by the contents of M between
the limits imposed by L and U. -M, L, and U are
locations within DDT-8 which may be opened,
modified and closed exactly as any general loca-
tion in the user's program.

Punch leader/ trailer tape

Punch binary tape from memory bounded by ad-
dresses mmmm and nnnn
Punch end-of-tape (i.e., checksum and trailer)

Accumulator storage (at breakpoint)

Lower limit of search

Upper limit of search

Mask; used in search

Link storage (at breakpoint)

NOTE: The character [ is generated by depressing ALT MODE on the key-
board.

For a thorough description of DDT-8, see DDT-8, Order No. DEC-
08-CDDA-D.

DEBUGGING WITH ODT-8

Using ODT-8, the programmer can run his binary program on the
computer, control its execution, and make alterations to his program
by typing on the Teletype keyboard. ODT-8 has the same capabilities
as DDT-8 except that the programmer must reference his program
using its octal representation instead of mnemonic symbols, and ODT-8
commands are formulated differently.

ODT-8 occupies less core memory than DDT-8, and can be loaded
into either lower (SA = 1000) or upper (SA = 7000) core memory,

Output
[T
mmmm;nnnn[P

[E
Address Tags

[A
[L
[U
[M
[Y

6-50

depending on where the user's program resides. That is, if the user pro-
gram resides in the first few pages bf memory, then ODT-8 should be
loaded in the upper pages of memory, and vice versa. As with DDT-8,
the user program can not occupy (overlay) any location used by
ODT-8, including the breakpoint location which is location 0004 on
page zero.

When the programmer is ready to start debugging a new program at
the computer, he should have at the console:

1 . The binary tape of the new program.

2. A complete octal/ symbolic program listing.

3. A binary tape of the ODT-8 program (either high or low
version).

To begin the debugging run, first ascertain that BIN is in core mem-
ory, then load the programmer's binary tape followed by the binary
ODT-8 tape (see Figure 6-4 for loading procedures). Figure 6-10
illustrates the procedures for loading and executing ODT-8.

(load Object Program)- See Figure 6-4

tZLT

Load ODT-8

See Figure 6-4

Set
SR = 1000 (low)
SR=7000(hi)

Depress LOAD ADD

Depress START

/Debug Object Programs
I ODT-8 Is In ■ J

y Command Mode y

Figure 6-14. Loading and Executing ODT-8
6-51

The example program debugged earlier in this chapter under "De-
bugging with DDT-8" could be debugged here. However, since the only
difference would be the appearance of the debugging commands, we
need only show the command formats and give a brief explanation of
each,, which can be found in Table 6-8.

Table 6-8, ODT-8 Commands

Command

nnnn/

RETURN

LINE FEED

t ( SHIFT/ N)

<- (SHIFT/O)

nnnnG

nnnnB
A

LINE FEED
C

nnnC

LINE FEED

nnnnW

Meaning

Reopen last named location; modify if desired
Open location nnnn; modify if desired
Close currently opened location (typing another com-
mand also closes the currently opened location)
Close currently opened location and open next se-
quential location

Close currently opened location, take contents as MRI
and open and type contents of referenced location
Close currently opened location and open indirectly,
i.e., the content of the opened location is interpreted
as the address of the location whose content is to be
typed and available for modification.
Go to location nnnn and transfer control to pro-
grammer for modification if desired.
Remove previously established breakpoint and re-
store breakpoint to original status.
Establish a breakpoint at location nnnn.
Open location containing the contents of the accumu-
lator (AC).

Open location containing the contents of the link.

Continue (proceed) from a breakpoint, i.e., execution

begins with the breakpoint instruction.

Continue (proceed) from a breakpoint and iterate

nnn (octal) times through breakpoint.

Open search mask, i.e., open for modification the

location containing the current value of the search

mask (mask is initially set to 7777) .

Open lower search limit (initially set to 0001; change

by typing new lower limit after ODT-8 has typed

0001).

Open upper search limit (initially set to SA of
ODT-8, 1000 or 7000; change by typing new upper
limit after ODT-8 has typed initial limit).
Word search, i.e., using the mask and lower and
upper limits, ODT-8 searches for instruction nnnn.

6-52

Table 6-8. ODT-8 Commands (Cont)

Command

Meaning

mmmm;nnnnP

Punch leader/ trailer tape, i.e., type T and then tHrn
punch ON; turn punch OFF when sufficient tape has
been punched, then depress STOP on the console. Re-
start ODT-8 at appropriate SA.
Punch binary core image of locations mmmm through
nnnn. After executing this command by typing the
RETURN key, the computer halts for the program-
mer to turn the punch ON and to depress CONT on
the console. When specified image is punched, turn
punch OFF.

Punch checksum and trailer tape. After executing this
command by typing the RETURN key, the computer
halts for the programmer to turn the punch ON and
to depress CONT on the console. When sufficient
trailer tape has been punched, depress STOP on the
console and restart ODT-8 at appropriate SA.

The address of the current or last location opened is remembered by
ODT-8 even after the commands G, C, B, T, E, and P, and may be
reopened merely by typing / (the slash command) .

After debugging his program, the programmer can command ODT-8
to produce a binary tape of the debugged program, which can be loaded
by the BIN Loader and run on the computer. The P or punch command
is used with the low-speed punch only. Figure 6-15 illustrates the pro-
cedure for generating the binary coded object tape using the high-speed
punch.

Typing the G command alone is an error, it must be preceded by an
octal address to which control will be transferred. If not preceded by
an address, a question mark will not be typed but it will cause control
to be transferred to absolute location 0.

Typing any pfhch command with the punch ON is an error and
will cause ASCII characters to be punched on the binary tape, which
means the tape cannot be loaded and run properly.
DEBUGGING NOTES. Only one breakpoint may be in effect at one
time, therefore, requesting a new breakpoint removes the previously
existing breakpoint. A breakpoint must not be set to any location in
the program which is altered during execution; if so set, the alteration
will destroy the breakpoint.

6-53

CWithODT-8 *\
In Control J

0—

Set 5R*6026

Depress D£P

Depress DEP

SR = 1225 (low)
SR* 7225(hi)

Depress LOAO ADD

Depress HSP ON

Leader Tope
Is Punched

Depress STOP

SR * 1203 (low)
SR * 7203{hi)

Type Command
mmmm;rtnnnP

Set
SR'!Z3Hlo«!
SR'7231lrii)

Depress LOAD ADD

Oepress LOAD ADD

Core Image Of
Is Punched

SR ■ !222(low)
SR= 7222 (hi!

Depress START

Accumulated
Checksum B Trailer
Tape Is Punched

Set
SR - 1231 (low)

SR = 7231(hJ)

Depress LOAD ADD

Depress HSP OFF
Remove Tape I

I Depress LOAD ADD

j Sft SR-6046

I Depress DEP

I S«l SR=604I I

Depress DEP

Sel
SR. 1000 (law)
SR.7000(hi)

| Depress LOAD ADD j

Depress START

^ OOT-8 Is Ready "\
^ For Next Command J

Figure 6-15. Generating Binary Tape Using High-Speed Punch

6-54

An open location can be modified by typing the desired octal instruc-
tion and closing the location. Any octal number from 1 to 4 digits in
length is a legal input, but typing a fifth digit is an error and will cause
the entire modification to be ignored and a question mark to be typed
by ODT-8.

ERROR DETECTION. The only legal inputs to ODT-8 are control
commands and octal digits. Any other characters will cause the charac-
ter or line to be ignored and a question mark (?) to be typed out by
ODT-8.

For a more detailed discussion, see ODT-8, Order No. DEC-08-
COCO-D.

FOCAL

FOCAL (Formula CALculator) is an online, conversational inter-
preter designed to help scientists, engineers, and students solve com-
plex numerical problems. The language consists of short easy to learn
imperative English statements. Mathematical expressions are typed in
standard notation.

With FOCAL, the programmer has the full calculating power and
speed of the computer at his fingertips. It is used for simulating mathe-
matical models, for curve plotting, for handling sets of simultaneous
equations in n-dimensional arrays, and many other kinds of problems.

The procedure for loading the binary-coded FOCAL tape is illus-
trated in Figure 6-16. When FOCAL is properly loaded in core, the
programmer sets the switch register to 0200 (the starting address of
FOCAL), depresses the LOAD ADDress switch and then the START
switch, and FOCAL will respond with its initial dialogue (see Chapter
9). The initial dialogue is a question and answer sequence, with
FOCAL asking questions and the user providing the answers.

After the programmer has answered the initial dialogue questions,
FOCAL will type an asterisk, indicating that it is ready to accept com-
mands from the programmer.

The FOCAL language and its numerous features are explained in
Chapter 9.

6-55

—

Load BIN

...

See Figure 6-3

Put FOCAL Tapt
In R«ad«r

Set SR-7777
And Press LOAD ADD

Turn TTY
To LINE

High -Spttd

-/Which ^>^
<T^ Reader \

Low- Speed

Set SR ■ 3777

- — "

Set LSR to START

* '■- :

Press CONT
1

Y«i

S«t SR-0200
And Pri LOAD ADD

Storting Addrtn

Of FOCAL

ftetpond To
Initial Dialogue

(FOCAL It Ready ""\
For User Input J

Figure 6-16. Loading and Executing FOCAL
6-56

Chapter 7

Disk
Monitor System

The PDP-8 Disk Monitor System is designed for any computer in
the PDP-8 family having at least one DECdisk. This system consists
of a keyboard-oriented Monitor, which enables the user to efficiently
control the flow of programs through his PDP-8, and a comprehensive
software package, which includes a FORTRAN Compiler, Program
Assembly Language (PAL-D), Edit program (Editor), Peripheral
Interchange Program (PIP), and Dynamic Debugging Technique
(DDT-D) program. Also provided is a program (Builder) for generat-
ing a customized monitor according to the user's particular machine
configuration (amount of core, number of disks or DECtapes, etc.).

The system is modular and open ended, permitting the user to con-
struct the software required in his environment, and allows the user full
access to his disk (or DECtape)— referred to as the system device—
for storage and retrieval of his programs. By typing appropriate com-
mands to the Monitor, the user can load a program (construct it from
one or more units of binary coding previously punched out on paper
tape or written on the disk by the Assembler, and assign it core),
save it (write it out, with an assigned starting address, on the system
device), and later call it (read it back into core from the system de-
vice) for execution.

7-1

GENERAL DESCRIPTION

The PDP-8 Disk Monitor System permits the user to control the flow
of programs through his computer and takes full advantage of the
extended memory capabilities of disk or DECtape. In addition to
the Monitor, the system also contains a library of system programs.
Together, they provide the user with the capabilities of compiling,
assembling, editing, loading, saving, calling, and debugging his own
programs.

Monitor Residence

Monitor, as well as system and user programs, is stored on and re-
trieved from the user's system device. To obtain a working Monitor,
the user must first build his own customized version, via the easy-to-
use dialogue technique of the System Builder program and store this
version on his system device. Following this, the user then creates his
System Program Library on the system device. Both of these procedures
are described in Appendix A of the Disk Monitor manual (DEC-D8-
SDAB-D).

In core, the resident part of Monitor (called head of monitor)
resides in the top page (locations 7600 through 7777) of field 0. The
starting address of Monitor is 7600; 7642 is the entry address to the sys-
tem I/O routine, which performs all reading and writing on the system
device. Nonresident portions of Monitor, such as those routines which
perform SAVEs and CALLs, are automatically called in as needed,
and in core they share the area from location 7000 through 7577.
(These portions disappear after use, leaving this area for the user.)

System Modes

At any point in time, the system is running in one of two modes:
Monitor mode or user mode. Monitor mode is entered (1) whenever
the Monitor is started (see "Initializing the Monitor") or (2) when
CTRL/C ( t C) is typed while running-any system program. Monitor
mode is signalled by the Monitor typeout of a dot ( • ) . At both Monitor
and system program time, Monitor is able to sense at C typein,
causing the system to enter Monitor mode, return to Monitor at loca-
tion 7600, and respond with a dot (•) typeout. At this point, the user
can issue any Monitor command via the Teletype keyboard.

User mode is present whenever the system is executing a system or
user program. System programs signal user mode by responding with
an asterisk (*) typeout.

7-2

SYSTEM PROGRAM LIBRARY

The Monitor System's library of programs presently consists of the
Peripheral Interchange Program (PIP), Disk System Editor (Editor),
PAL-D Disk Assembler (PAL-D), 4K Disk FORTRAN (FORTRAN
D), and Dynamic Debugging Technique for Disk (DDT-D), and this
list is destined to lengthen with time.

To load a program using the Monitor System, the Loader makes
certain queries to which the user must type a reply. The queries are
the same for all programs. The user's replies will vary, however, de-
pending on the particulars of the program being loaded.

When loading a program into core, the user should first check to see
whether Monitor is in core. This is done by typing t C (CTRL key
and then the C key). The tC will not echo (not print on the tele-
printer). If Monitor is in core, it will respond by typing a period (•)
at the left margin of the teleprinter paper. If a period is not typed in
response to t C, Monitor is not in core.

The library system includes the Disk System Binary Loader
(LOAD) which is automatically saved on the disk at Build time.
Disk System Binary Loader is described in "Loading Programs — Disk
System Binary Loader."

The user may save any program on the disk by responding to the
last period typed by Monitor with the word SAVE,, a l-to-4-character
name of the program, the type of program (user or system), whether
it's a one or more page save, and the location of its starting address,
as is thoroughly described in "Saving Programs (SAVE Command)."

After each program is saved on the system device, it may be called
(i.e., transferred from the disk into core) merely by responding to
Monitor (to a period) with the four characters designated as the name
of that program, as explained in "Calling a Program (CALL Com-
mand)."

The programs in the System Library are described below. Further
information concerning these programs may be found in the Disk
Monitor manual or in the other manuals referred to.

Peripheral Interchange Program

The Peripheral Interchange Program (PIP) performs general utility
operations, such as listing the contents of specified directories, deleting
unwanted files from the system device, and transferring files between
devices, and copying specified files. PIP enables the user to do any of
the above operations merely by typing commands from the teleprinter
keyboard.

7-3

A summary of PIP options follows.

RETURN or A

List entire system directory
Copy a binary file
Delete a file to be specified
Copy a FORTRAN binary file
Move directory to safe disk
Protect disk 1 (blocks 0-176)
Restore directory from safe disk
Copy a system file 1
Copy a user file 1
Copy a ASCII file

Disk System Editor

The Disk System Editor enables the user to generate and edit
symbolic programs online from the teleprinter keyboard. The sym-
bolic program may be either printed on the teleprinter, punched
on paper tape using the high- or low-speed punch, or stored on the
system device as a user program.

Editor operates either in command or text mode. In command
mode, all typed input is interpreted as a command instructing Editor
to perform a certain operation or to allow the user to perform an
operation on the text stored in the buffer. In text mode, all typed
input is interpreted as text to replace, to be inserted into, or to be
appended to the contents of the text buffer.

The command language of the Disk System Editor is identical to
that described in the PDP-8 Symbolic Editor manual (DEC-08-
ESAB-D) with a few exceptions.

A summary of Editor commands follows. Each command is termi-
nated by pressing the RETURN key, which initiates execution of
the command.

Function

Command

Meaning

Read

Read incoming text and append to buffer until a
form feed is encountered

Append

Append incoming text to any already in the buffer
until a form feed is encountered.

List

m,nL

List the entire buffer.

List the line n.

List lines m through n.

7-4

Function

Command

Meaning

Proceed

Proceed and output the entire contents of the
buffer and return to command mode.

Output line n, followed by a form feed.

m,nP

Output lines m through n, followed by a form feed.

Trailer

Punch four inches of trailer.

Punch the entire buffer and a form feed; kill the
buffer and read next page.

Repeat the above sequence n times.

Kill

Kill the buffer.

Delete

Delete line n.

m,nD

Delete lines m through n.

Insert

Insert before line 1 all text until a form feed is
encountered.

Insert before line n until a form feed is encount-
ered.

Change

Delete line n and replace it with any number of
lines from the keyboard until a form feed is en-
countered.

m,nC

Delete lines m through n, replace from keyboard
as above until form feed is encountered.

Move

m,n$kM

Move and insert lines m through n before line k.

Get

Get and list the next line beginning with a tag.

Get and list the next line after line n which begins
with a tag.

Search the entire buffer for the character specified
(but not echoed) after the carriage return;
allow modification when found.

Search line n, as above, allow modification.

m,nS

Search lines m through, n, allow modification.

End file

Process the entire file (perform enough NEXT
commands to pass the remaining input to the
output file) and create an end-of-file indication;
legal only for output to the system device. If
the low-speed paper tape reader is used for
input while performing an E command, the
paper tape reader will eventually run out of
tape, and at this point typing a form feed will
allow the command to be completed.

7-5

PAL-D Disk Assembler

PAL-D, the acronym for Program Assembly Language for the
Disk system, is the symbolic assembly program designed primarily
for the 4K PDP-8 family of computers with disk or DECtape.

The PAL-D Assembler performs many useful functions, making
machine language programming easier, faster, and more efficient.
Basically, the Assembler processes the user's source program state-
ments by translating mnemonic operation codes into the binary codes
needed in machine instructions, relating symbols to numeric values,
assigning absolute core addresses for program instructions and data,
and preparing an output listing of the program which includes
notification of any errors detected during the assembly process.

The Assembler is thoroughly documented in the PAL-D Disk
Assembler Programming Manual (Doc. No. DEC-D8-ASAA-D).

FORTRAN-D for the Disk System

FORTRAN-D (FORmula TRANslation for the Disk System),
is an expanded version of standard PDP-8 FORTRAN designed
for PDP-8 family computers with disk or DECtape units. FOR-
TRAN-D is described in detail in the 4K FORTRAN Programming
Manual (DEC-08-AFAC-D).

FORTRAN-D contains a compiler, an operating system, and two
programs to aid debugging. The FORTRAN compiler is used to
convert a source program into an object program. The FORTRAN
operating system is used to execute the object program.

This version of FORTRAN is designed to facilitate user/system
communication by allowing the user to type appropriate commands
from the teleprinter keyboard, eliminating the need to toggle input
using the switch registers.

■FORTRAN statements specify the computations required to carry
out the processes of the FORTRAN program. There are four types
of statements provided for by the FORTRAN language:

a. Arithmetic statements define a numerical calculation.

b. Control statements determine the sequence of operation in
the program.

c. Specification statements define the properties of variables,
functions, and arrays appearing in the source program. They
also enable the user to control storage allocation.

d. Input/output statements are used to transmit information be-
tween the computer and related input/output devices.

The compiler consists of a loader (FORT) and the main portion of
the compiler (.FT.). This version of the compiler differs from the

7-6

standard PDP-8 4K FORTRAN compiler in the following ways.

a. It uses the disk or DECtape unit (whichever is the system
device) during its operation,

b. It will compile programs which have been stored on the sys-
tem device or programs which have been prepared on punched
paper tape.

c. It will generate a FORTRAN binary output file either on the
system device or on punched paper tape.

d. Significant improvements have been employed with the READ
and WRITE statements.

e. Input and output devices are determined using the Command
Decoder.

f. It is possible to terminate compilation at any time by typing
t C, thus returning control to Monitor.

g. Within certain restrictions, a program compiled with output to
the system device may be executed immediately when the user
types f P after compilation of the program.

. h. Statement numbers need riot be delimited by a semicolon, un-
less the user wishes them to be employed for appearance,
i. Statements without preceding numbers must be preceded by a
space, a tab, or a semicolon.

Debugging Aid (Symbolprint). Symbolprint (STBL) is a program
which may be used with the FORTRAN compiler. Its purpose is to
help the user create and debug his FORTRAN programs by providing
certain information about the compiler-generated interpretive code.
Symbolprint may be used only immediately after a program has been
compiled using the Disk FORTRAN compiler.
Symbolprint provides the following information:

a. The core limits of the interpretive code.

b. A list of variable names and their corresponding locations
(the symbol table).

c. A list of statement numbers and their corresponding locations
(the statement number table) .

Operating System. The FORTRAN operating system consists, of a
loader (FOSL) .and the interpreter and arithmetic subroutine package
(.OS.). This version of FOSL differs from the paper tape FORTRAN
operating system in the following ways.

a. It will load and execute programs which have been compiled

and saved on the system device or programs which have been

compiled on paper tape.

7-7

b. FOSL may be called directly by the compiler when a program
has been compiled and saved on the system device. This is
referred to as compile-and-go mode.

c. FOSL is able to recognize READ and WRITE statements
which may read and write data in ASCII format on either
the low-speed paper tape reader/punch, the high-speed paper
tape reader/punch, or the system device.

d. The execution of a FORTRAN program may be interrupted
by the user at any time by typing T C; control will be returned
to Monitor.

Debugging Aid (Diagnose). Diagnose (DIAG) is a basic system pro-
gram which helps the user debug his FORTRAN program. It is in-
tended to be used in conjunction with the PDP-8 4K FORTRAN
Operating System and revised FORTRAN Symbolprint. Diagnose pro-
vides the following information.

a. If stack overflow or underflow has occurred, it will type a
message indicating which of the five run-time stacks caused
the error.

b. It will type a message indicating the contents of the current
location counter (CLC).

c. If tiie counter stack is nonempty, it will type the contents of
that stack.

d. If location zero is nonzero, it will type the contents of that
location (minus one), indicating the point at which some
FORTRAN systems error occurred.

Dynamic Debugging Technique for the Disk System

The Dynamic Debugging Technique for the Disk System (DDT-D)
is used for online checking, testing, and altering object programs by
typing from the teleprinter keyboard. When debugging online, the user
checks his program at the computer, controlling its execution, and
making corrections or changes to his program while it is running on
the computer.

When using DDT-D, the user should have a listing of his program
and its symbols so that he can update the program listing as correc-
tions and changes are made to his program. The user may refer to .
variables and tags by their symbolic names or by their octal values.

DDT-D operates as described in DDT Programming Manual (Doc.
No. DEC-08-CDDA-D), except where that manual differs from this
one, in which case this manual has precedence.

7-8

DDT-D can be considered as being in three sections.

a. DDT Proper (.DDT) A slightly modified version of

DDT-8 (low); occupies core loca-
tions 200 through 4577 and the
three breakpoint locations.

b. Driver (DDT) Resident in core with its origin set

above DDT proper (above 4577);
it is a 2-page program plus a 1-
page once-only program, and it
contains breakpoint insertion and
removal logic, overlay routines,
continuation iteration count and
control, and breakpoint list.

c. User Core Image File Occupies same storage area as

(.SYM) DDT proper and is used for swap-

ping DDT proper and the user
program to and from the system
device.
DDT-D is an expanded version of DDT-8 with the following excep-
tions.

a. Three breakpoints (as opposed to only one in DDT-8)

b. No punching (program may be output on the system device)

c. No switch options (user direction is via keyboard)

d. No halts (continues when user types t P)

EQUIPMENT REQUIREMENTS

The minimum equipment requirements of the PDP-8 Disk Monitor
System are as follows.

A basic PDP-8 family computer with 4K of core

Teletype

3-Cycle Data Break (Option required with PDP-8/S)

At least one DF32 Random Access DECdisk File or a TC01

DECtape Control with a TU55 DECtape transport. The DEC-

tape must have timing and mark tracks written on it prior to use.

NOTE

The system will recognize up to 32K of core, up to four disks (1 Type DF32
and 3 Type DS32's), up to eight DECtape transports (TC01 Control only) and
a high-speed paper-tape reader/ punch.

7-9

INITIALIZING THE MONITOR

The following discussion assumes that the user has built a customized
Monitor via the System Builder and has stored it on his system device;
it further assumes that the user has bootstrapped this customized Moni-
tor into core and that the Monitor is currently intact. If this is not the
case, such a Monitor must be built and bootstrapped into core bv the
steps outlined in the Disk Monitor System manual (DEC-D8-SDAB-

D).

Monitor start is at location 7600. A jump to this location can be
made by either ( 1 ) stopping the machine, setting the switches to 7600,
and pressing LOAD ADD (load address) and START, or (2) typing
t C when in Monitor mode or when a system program (or any user
program which includes coding to sense atC typein) is running. 1
Monitor start performs the following actions.

a. Saves the coding from location 7200 through 7577 in the first
two scratch blocks on the system device.

b. Reads blocks 1 and 2 (containing the rest of Monitor) from
the system device into these locations.

c. Transfers control to Monitor, which responds with a carriage
return, line feed, and a dot.

A Monitor restart can be performed by typing RUBOUT to Monitor.
A Monitor restart performs the same actions as described above except
for Subparagraph a. A common use for RUBOUT is to terminate a
command string when the operator has discovered that he has made a
mistake. The command string is ignored, and Monitor responds as
described in Subparagraph c. The user core image on the system de-
vice is not changed by RUBOUT (it is changed, however, by fC).

COMMAND STRINGS

The user types commands in the form of command strings to direct
Monitor, or a system program, to perform some action. Command
strings are simple in format and afford the user an easy means of com-
municating with the system.

Monitor indicates its readiness to accept a command string by typing
a dot, and at this point, the user can type some Monitor command,
such as CALL or SAVE.

Unless otherwise indicated, all command strings are terminated by
the RETURN key, which echoes as a carriage return, line feed.

*A start instruction (ST=7600) is issued when running Loader causes a jump to
7600 after loading has been performed. Certain errors also cause a jump to this
location.

7-10

System programs indicate their readiness to receive information by
typing either an asterisk or a query. The most common queries are as

follows.

*OUT- Requests that the user specify one output device name. In
the case of disk or DECtape the filename to be assigned to
the output data must also be specified (see "Filenames").
*TN- Requests that the user specify one or more (up to five)
input device names. For disk and DECtape, filenames of
input files must also be specified (see Filenames) .
*OPT- Requests that the user specify one option or switch, entered
as a single alphanumeric character; see the appropriate
reference for options available in each system program.
This communication between the system and the user is handled by
a portion of Monitor known as the Command Decoder. 1 Command
Decoder is called into core by the system when needed and occupies
any four contiguous pages of core.
Command String Format

Command strings are composed of a few basic elements and follow
certain rules of punctuation. Their basic elements are as follows.

a. Device names

b. Filenames

c. Punctuation

d. Special characters

Each of these elements is described in the following paragraphs.

Device Names. Device names permitted in command strings are as
follows.

Dn: DECtape unit, if both disk and DECtape are present in the sys-
tem (n=unit number, through 7)
S: System device (disk or DECtape unit 0)
R: High-speed paper tape equipment (reader or punch)
T: Low-speed paper tape equipment on the Teletype (reader or
punch)

Filenames. Filenames are limited to four characters in length and can
be composed of any combination of alphanumeric characters or special
characters 2 with the following exceptions.

a. Imbedded spaces cannot appear in a filename (they are
ignored). 3 However, trailing spaces are permitted.
iCommand Decoder is a system program (.CD.) which is saved on the system
device at Build time.

^Although both printing and nonprinting keyboard characters are allowable,
printing characters are recommended.

3 Note that Monitor is given the filename EX C; one reason for this unconven-
tional use of an imbedded blank is to protect Monitor from accidental destruction
by the user (e.g., deletion via PIP).

7-11

b. A filename cannot be one of the following words or symbols.
CALL SAVE ! , ; :
Extensions to the filenames specified by the user are automatically
appended by the system. They are used internally by the system and
cannot be referred to or modified by the user.

SYS (n) Saved system program file in core bank n.

USER (n) Saved user program file in core bank n.

ASCII Source language program file (input to PAL-D Assembler

or FORTRAN Compiler).

BINARY Binary program file (output from PAL-D Assembler).

FTC BIN Interpretive binary file (output from FORTRAN Com-
piler) .

Filenames (and extensions) are meaningful only for file-structured
devices (disk and DECtape). If they are specified for other devices,
they are ignored. Both the filename and extension name appear on
directory listings produced by the list feature in PIP. 1

Example:

NAME

TYPE

BLK

PIP ,

, SYS (0)

0015

SRC1

. ASCII

0007

BIN .

BINARY

0001

SRC1 . USER (0) 0001

Punctuation. Punctuation within command strings is as follows.

, Used to separate device names, when more than one is given in a

command string. The comma is also used to separate core references

in a SAVE command string, when more than one contiguous area

of core is specified.
; Precedes the entry specification in a SAVE command.
: Terminates each device name. The colon is also used following the

filename in a SAVE command to indicate that the file is to be saved

as a user program.
- Separates the beginning and ending addresses of a contiguous core

area specification in a SAVE command.
! Follows the filename in a SAVE command when a file is to be saved

as a system program.
t"8D" in example means VERSION 8, change D.

7-12

Special Characters. Special characters are used as described below,
t C If given while the system is in Monitor mode or a system

program is running, control is returned to Monitor start
(location 7600). Monitor responds with a dot. f C is typed
by holding down the CTRL key and striking f C. t C does
not echo (does not print).
fP Typed in response to a f typeout. Instructs the system to

proceed with the next operation. 1 f P is typed by holding
down the CTRL key and striking P. f P does not echo
(does not print).
RETURN Carriage return terminates current command string input.
When typed alone, in response to a system query, it indi-
cates that the user does not desire to specify the item (e.g.,
device name) requested. Echoes as carriage return, line
feed.
RUBOUT Causes the current command string to be ignored, and the
system returns to the beginning of the command string and
is ready to receive a new command. RUBOUT does not
echo.
Examples of Command Strings

These examples illustrate the elements and rules explained above.
Samples of both Monitor commands and system program commands
are given. 1 *

Monitor Commands

. CALL PRG1 Call the user program file, PRG1,

from the system device into core for
execution.
. SAVE PALD! 0-7577; 6200 Save a program, previously loaded by

Loader into locations through 7577
or core, on the system device as a
system program (!). Assign a starting
address of 6200 and a filename,
PALD.

SYSTEM PROGRAM

COMMANDS /

*IN- S:PRQ2 Use the file PR02 on the system de-

vice as the input file.

*IN-S:TST1,R: Use the file TST1 on the system de-

vise and one file from the high-speed
paper tape reader as the input files.

*OUT- D5:SPEC Write the output file on DECtape

unit No. 5 and assign it the filename
SPEC.

x \ P can also be used to prematurely terminate certain operations while ^m pro-
gress (e.g., the typing out of a file directory by the list option in PIP).
2 In all examples, system response (typeout) is underlined for clarity.

. 7-13

LOADING PROGRAMS— DISK SYSTEM BINARY LOADER'

The Disk System Binary Loader takes as input the binary coding
produced by the PAL-D Assembler and loads it into core in executable
form. When loading is completed, Loader "disappears" after first en-
tering the loaded program at the starting address typed by the user
just prior to loading (see "Loader Operating Procedures" below).
Loader accepts input from the system device or paper tape.

Loader requires one pass for any program which does not load above
location 6777 (field 0). Loader uses core from location 167 through
177 and 6000 through 7577, and the resident portion of Monitor occu-
pies the remainder of field 0. One-pass loading reads input files only
once.

Two passes are required for all other programs (i.e., programs load-
ing above 6777). In two-pass loading, programs can be loaded in all
of field 0, except locations 7600 through 7777. 2 Two^pass loading re-
quires that input paper tapes be read through the reader twice.

Loader Operating Procedures
. LOAD

•IN-

Direct Monitor to bring Loader from
the system device into core for exe-
cution.

Loader requests source of input (s).
Type one or more device names,
separated by commas. If an input
device is a file-structured device (S:,
Dn:) include filename(s).
Up to five files can be specified.

Examples
*IN-R:,R:,R:

*IN-S:INPT

*IN-S:BIN2,R:

*IN-S:BIN1,S:BIN2

Input three tapes from the paper

tape reader.

Input the file INPT from the system

device.

Input the file BIN2 from the system

device and one tape from the paper

tape reader.

Input the files BIN1 and BIN2 from
the system device.

x The Disk System Binary Loader is a system program saved on the disk at Build
time. It is called by the user in the same manner as any system program. It
occupies locations 7000-7577 and has a starting address of 7000.
-In 8K and larger systems, Loader sets up locations 7574 through 7577 to per-
form a start in fields other than 0. It is the user's responsibility to. protect these
locations if he wants to start in other than field 0.

7-14

*OPT-

Examples
*OPT-l

*OPT-2

(or anything else)

*ST =

Examples
*ST =

*ST =

7600

*ST =

30225

*ST =

10000

If device(s) are valid and filenames
(if any) are actually found on the
system device, Loader responds with
one asterisk for each correct input.
Loader requests mode desired (one-
pass or two-pass).

One-pass loading desired; no pro-
grams are loaded above location
6777.

Two-pass loading desired; programs
can be loaded above location 6777.

Loader requests the starting address
to which control is to be transferred
when loading is completed. The ad-
dress is typed in the form

fnnnn
where

f = field number 1 (omitted if

field 0).
and

nnnn = location within field

Load into field 0.

Return to Monitor after loading.

Load into field 3.
Jump to location 255. field 3, after
loading.

Load into field 1.
Return to Monitor after loading
into field I.

Loader now types a series of up-
arrows, one at a time, as explained
below.

Following each up-arrow typeout,
the user is required to perform
one or more actions.
tttf First up-arrow: Loader is ready to

load. If paper tape input, put the
tape in the reader. Type t P-"

'The f-digit forces Loader to start loading into the specified field until a "field

setting" is found in the input file or tape.

2 Jf Teletype paper tape equipment is used, type t P hefore turning on the

reader.

7-15

Second up-arrow: End of pass 1. If
operating in one-pass mode, type
t P to jump to previously specified
starting address.

If operating in two-pass mode,
type? P.
The next two up-iarrows appear only
if operating in two-pass mode.
Third up-arrow: Reload paper tape

input for pass 2. Type fP.
Fourth up-arrow: End of pass 2.
Type f P to jump to previously
specified starting address.
Multiple Input Files
An up-arrow is typed out as the
processing of each input file is com-
pleted. If paper tape input, insert the
next file in the reader and type fP.
Repeat the above step until all files
given in response to the *IN-request
have been processed.
If in two-pass mode, each tape must
be entered twice, in the order

T1,T2, T3 T1.T2, T3,

After all files have been entered the
required number of times, type f P
to jump to the previously specified
starting address.

NOTE

After each input paper tape is read, the high-speed paper tape version of
Loader loops until the user types. fP to continue. However, the low-speed paper
tape version halls. Thus, when using the Teletype paper tape equipment for
input, the user need not type f P but press CONT on the console and start the
paper tape reader.

At this point, Loader disappears and control is transferred to the
previously specified starting address.

The loading parameters for system programs are given in Table 7-1.

Loader Error Messages

An illegal checksum error condition causes Loader to type

and return to Monitor after the user types tP or fC. Error messages
for illegal filenames or devices are as specified in "System Error

Messages."

7-16

SAVING PROGRAMS (SAVE COMMAND)

The SAVE command enables the user to write core images of sys-
tem or user programs from core onto his system device for subsequent
call-in (CALL) and execution. For example, a program which has
been loaded by Disk System Binary Loader can be stored on the sys-
tem device by the SAVE command. Or, a previously saved program
which has been called in and modified by DDT can be stored
in its updated version on the system device, overlaying the old version
if desired.

Core images can be saved in units of one or more pages, each page
occupying one block on the system device. If a core specification (see
below) addresses only a portion of a page, the entire page is written
out. For example, the core specification 45-150 is treated as though it
were 0-1 77-. Core areas to be saved may be contiguous or noncon-
tiguous as desired by the user. Up to 32 1( , core specifications, in any
combination of monotonically increasing single-page or multiple-page
requests, can be entered in a single SAVE command.

SAVE Command Format

SAVE filename

SAVE
filename

or :

1
J

core-specifications, , . .; entry-point

Directs Monitor to call in the nonresident SAVE
routine.

The filename (program name) to be assigned to
the file on the systems device. This name will be
used to call the file later when the user wants to
read in and execute the program. Restrictions on
the formation of filenames can be found in
"Command Strings" under "Filenames." Any pre-
viously saved program with the same "filename"
and having the same extension will be automatic-
ally overwritten.

! is typed immediately after the filename of a file
if the user desires to save it as a system program
(e.g., PIP). A program saved in this manner can
be called in by simply typing its name to Monitor
(the word CALL is not required)

_;filename
An extension name of .SYS is automatically ap-
pended to the filename.

7-17

: is typed immediately after the filename of a file
if the user desires to save it as a user program. A
program saved in this manner can be called in
and executed later via the CALL command

^CALL filename
An extension name of .USER is automatically
appended to the filename,
core-specifications Up to 32 specifications can be entered in a single
SAVE command. Each core specification is sep-
arated from the following one by a comma. The
last core specification in the series is followed by a
semicolon. Addresses are expressed in octal.

Single-page core specification

fnnnn
where

f= field number (can be omitted if field
0).
nnnn— any location within the page which
the user desires to save.

Multiple-page core specification
When a user wishes to save a core area of several
contiguous pages, he can type a multiple-page core
specification in the format.

fnnnn,-nnnn.,
where

f=field number (can be omitted if field
0).
nnnn, = any location within the first page of
the series of contiguous pages to be
saved.
nnnn 2 =any location within the last page of
the series of contiguous pages to be
saved.
The following rules apply.

a. The beginning address of a multiple-page re-
quest must be smaller than the ending address
(nnnn t must be smaller than nnnn 2 ).

b. Both addresses must be in the same field.

c. The field number (f) must be within the
range of your system; however, no check for the
validity of this number is performed at SAVE
time.

7-18

entry-point The entry point of the saved program, in the for-

mat

fnnnn (see explanations above)
An entry point of causes a return to Monitor at
■ CALL time, regardless of the field into which the
program was saved.

NOTE

The last nonzero field number encountered in a SAVE command string is re-
membered and prefixed to all other addresses in the command string. (Remem-
ber: only one field can be referred to in each command string.)

Example: The following entries are identical in meaning.

SAVE PRGA: 10000-10777, 11400, 1600-17777; 10200

SAVE PRGA: 30000-777, 51400, 26000-7777; 10200

SAVE PRGA: 10000-777,1400, 6000-7777; 200

SAVE PRGA: 0-777, 1400, 6000-7777; 10200

In each of these examples, all addresses are treated as being field I, because the

last five-digit entry seen contained a most significant digit 1.

SAVE command strings for system programs are given in Table 7-2.

SAVE Command Processing

A list of the required pages is constructed from the information
typed by the user and a block requirement count is kept. When the
user types the terminating carriage return (RETURN), allowing the
SAVE process to begin, a directory name search on the system device
is initiated. If a file having the same name as the filename in the SAVE
command is found, it is replaced by the file now being saved. If no such
file is found, a new file is created. Next, a storage availability search
finds a sufficient number of available blocks on the system device to
satisfy the block requirement count. (See above.) These block numbers
are stored in a corresponding block list; the blocks are then filled with
the contents of the pages to be saved. When the SAVE process is com-
pleted, control returns to Monitor (7600).

CALLING A PROGRAM (CALL COMMAND)

Once a file has been loaded and saved, it can be called into core as
desired. There are two types of CALL command strings: one for system
programs and the other for user programs.

The CALL command string format for system programs (programs
saved by a SAVE command string in which the filename was followed
by a !) is

.filename
where filename is the same as the one used in the SAVE command
string which saved it.

7-19

The CALL command string format for user programs (programs
saved by a SAVE command string in which the filename was followed
by a : ) is

JCALL filename

When a program is called, a directory name search is performed on
the system device. Associated with the directory entry is the entry point
of the program and information concerning file protection and memory
extension. If the appropriate directory name entry is found and the
file has the proper extension (.SYS or .USER), calling proceeds. If
not, the calling process is terminated, ? is typed, and control is returned
to Monitor.

Table 7-1. Loading Parameters for System Programs

Name

Core Limits Entry Point

Pass

PIP

0-5177 1000

EDIT

0-3177 2600

PALD

0-3377, 3600-4377, 6200
4600, 5200,
6200-6577, 7000-7577

FORT

0-1777 200

.FT.

200-7377

STBL

600-777 600

FOSL

0-1577 200

.OS.

0-5177

DIAG

200-1177 200

.DDT

200-4577

.SYM

200-4577

DDT

7200-7577 7200

Table 7-2. SAVE Commands for System Programs

Name

SAVE Command String

PIP

SAVE P1P!0-5177;1000

EDIT

SAVE EDIT!0-3177;2600

PALD

SAVE PALD!0-7577;6200

FORT

SAVE FORT!0-1777;200

FT.

SAVE FT.!200-7377;0

STBL

SAVE STBL!600;600

FOSL •

SAVE FOSL!0-1577;200

.OS.

SAVE ,OS.!0-5177;0

DIAG

SAVE DIAG!200-1177;200

.DDT

SAVE .DDT!200-4577;0

.SYM

SAVE .SYM!200-4577;0

DDT

SAVE DDT!7200-7577;0

(User may assemble anywhere above location 4577)

7-20

SYSTEM ERROR MESSAGES

As an input command string is being typed, Monitor recognizes any
incorrect syntax and remembers it. When the user types a carriage re-
turn, Monitor responds with a ? to indicate invalid input.

Error messages output by Command Decoder arc given in Table 7-3.

Table 7-3. System Error Messages

Message Meaning

? Illegal syntax or miscellaneous error condition

D Directory on the systems device is full

E Too many inputs or outputs were entered

I No such inputs

S . System I/O failure

Monitor-time read or write errors cause a halt to occur. Persistence
of this condition indicates a hardware failure, as the system I/O rou-
tine attempts to read or write three times before halting.

I/O PROGRAMMING

The modular concept of input/output (I/O) handling of the disk
system provides for easy maintenance and programming. The system
device I/O is found in the following places (all I/O routines must be
infield 0).

a. Top page of field (location 7642) which is the I/O routine
used by all system programs for normal I/O. A copy of this
page is on block of the system device. Block of each DEC-
tape is the DECtape I/O routine.

b. Interrupt versions of disk and DECtape routines are found in
PIP.

c. Paper tape I/O is handled by individual programs.

The basic I/O routine (see "Normal," below) is called as shown in
Table 7-4. It is called in two ways, as determined by bit 2 of the func-
tion word.

Normal. The I/O routine returns to JMS +6 (normal) or JMS +5
(error). For example, the following routine would read consecutive
blocks from a file on the system device. The routine is initialized by
putting the first block number of the desired file into location LINK.

7-21

If an attempt is made to read past the last block of the file, an exit
will be made to a routine called ENDFIL.

GETBLK,

TAD LINK

/GET LINK FROM LAST READ

SNA

/IS THIS END OF FILE

JMP ENDFIL

/YES

DCA BLOK

JMS 7642

/CALL DISK I/O ROUTINE

BUFFAD

/FUNCTION=READ

/BUFFER ADDRESS

LINK,

JMP ERROR

/ERROR RETURN

JMP I GETBLK

Indirect. The I/O routine returns to the 12-bit address in the error re-
turn word (normal or the 12-bit address in ERROR ).

Table 7-4 Calling Sequence for Disk System I/O Routine

Calling Sequence

Explanation

JMS I SYSIO

Location SYSIO points to I/O

FUNCT

Function word 1

BLOCK

Block to be accessed

CORE

Low-order core address

LINK

Filled by READ, used by WRITE

ERROR

Error return here
Normal return here

function word: Bits 0-1

unused

Bit 2

=0,

normal return -

= 1,

indirect return at end of read/ write to
address +1 in error return

Bits 3-5

unit

no. if DECtape

Bits 6-8

memory field

Bits 9-1 1

function: READ = 3; WRITE = 5

7-22

Chapter 8

Time-Sharing
System

INTRODUCTION

The Time-Sharing System for the PDP-8/I and PDP-8 computers
(TSS/8) is a general purpose stand-alone time-sharing system offering
each TSS/8 user a comprehensive library of system, service, and utility
programs. These programs provide facilities for compiling, assembling,
editing, loading, saving, calling, and debugging user programs online.
Also included are two conversational language programs, FOCAL-8
and BASIC-8. 1

By segregating the central processing operations from the time-con-
suming interactions with the human users, the computer can in effect
work on a number of programs simultaneously. Giving only a fraction
of a second at a time to each program or task, the computer can deal
with many users seemingly at once, as if each user has the computer to
himself. The executions of various programs are interspersed without
interfering with one another and without detectable delays in the re-
sponses to the individual users.

Core and Disk Allocation

TSS/8 requires a minimum of 12K words (three fields) of core mem-
ory. The first 8K (Fields and 1 ) are shared by the various Monitor
subprograms. The other 4K (Field 2) and any additional core memory
are shared by the users of the system.

i BASIC-8 is a modified version of the algebraic language originally developed
at Dartmouth College.

8-1

The disk area is divided as follows.

Monitor Area The first 16K of the disk is occupied by the

Monitor subprograms. These subprograms
are swapped into core memory as needed.

User Swapping Area This area consists of a 4K track for each
user in the system. When a user is tempo-
rarily swapped out of core memory to allow
other users to have their turn, his program
is stored on his 4K track.

File Storage Area The remaining disk area is used for storing
system and user program files.

Monitor Functions

The heart of this time-sharing system is a complex of subprograms
called Monitor. Monitor , coordinates the operations of the various
programs and Teletype consoles, allocates the time and services of the
computer to users, and controls their access to the system. The sched-
uling function includes scheduling the user's requests, transferring con-
trol of the central processor from one user to another, moving (swap-
ping) programs between core memory and disk, and managing the
user's private files.

TSS/8 User and Console

A TSS/8 user is any person running a computer program within
TSS/8. He has an account number and password assigned to him which
identifies him to TSS/8. This account number and password are assigned
to the user by the person responsible for the system, and it identifies
his files on a permanent basis. When the user is using the system, this
number also identifies his console, his individual disk swapping track
(a contiguous 4K section of disk in the User Swapping Area) for tem-
porary storage of active programs by the Monitor, and whatever other
facilities he may be using at any given time.

While the user is logged into the system, he owns at least one Tele-
type console (see Ghapter 4) and one 4K disk track. The console con-
sists of a keyboard, which allows the user to type information to his
user programs and to Monitor, a paper tape reader and punch for
paper tape input to and output from core, and a teleprinter, which fur-
nishes typed copy of user input and program and Monitor output.

8-2

System Program Library

A comprehensive library of programs is available to all TSS/8 users
and is located in the File Storage Area of the disk. The library can
consist of any or all of the system software which operates in 4K of core
memory. Any program in the library can be called into core from the
disk and started merely by typing, in response to the dot typed by
Monitor, the command R and the assigned name of the program. For
example,

.R FOCAL
brings FOCAL into core from the disk and automatically executes it so
that it begins typing out its initial dialogue.

User Programs

When a user program is being run by TSS/8 it is swapped into core
memory from the user's disk track. Several programs may be run at
virtually the same time by employing the technique of bringing a pro-
gram into core from the disk, allowing it to execute for a short time,
marking the state in which its execution is stopped, returning it to the
disk, and picking up the next user program.

User programs are serviced regularly on a round-robin basis. After
a user program has been executed, it is placed last in the queue of user
programs waiting to run. Each program is allowed to run a fixed inter-
val of time and then it is exchanged for the next user program. If only
one program is in a condition to run, it is allowed to run without inter-
ruption.

User Files

User files are stored in the File Storage Area of the disk. Using the
appropriate Monitor commands, the user can create new files and ex-
tend, contract, and delete old files. Files may contain textual informa-
tion, binary core images, or data in any standard format.

Each user can have access to up to four active files at any time. An
internal file number (0 through 3) is associated with each of the user's
active files, and commands then operate in terms of the internal file
numbers. With this feature, a user may attach one of his files to an
internal file number and then load service and utility programs that
operate on his file without having to explicitly call the file for each
service or utility program used.

The user can protect his files against unauthorized access. He can
also specify the extent of access certain other users may have to his
files. For example, a user's associates may be permitted to look at the
data of certain files but not permitted to alter that data.

8-3

System Configuration

Depending on the hardware configuration of a particular TSS/8,
there can be from 1 to 16 users working with the system simultane-
ously. Each user has the following resources: at least 4K words of core
memory for execution of programs and a corresponding 4K disk track
for temporary storage of his core image when swapped out by Monitor.

The minimum equipment requirements of the system are listed below
(see Figure 8-1).

PDP-8/I or PDP-8 with KT08/I Time-Sharing Modifications

MC8/I-A Memory Extension Control and 4096 words

MM8/IA 4096 word memory

RF08 Disk Control

RS08 Disk

PT08 Asynchronous Line Interfaces, Dual (4), Real-Time Clock

PT8/I High-Speed Paper Tape Reader

KE8/I Automatic Multiply-Divider

H961A Option Cabinet

CONSOLES
CI TO 16 )

i, A M

SLUT

(TELETYPE MODEL 33 or 35)

CONSOLE CONTROL
DC08

or -
PT08's

PDP-8/1 (modified)

CPU

4K CORE

TT ADDITIONAL
I CORE
. (MAXIMUM
'CORE STORAGE
1 - 32K

J_ WC-RD§2 _

Figure 8-1 . System Configuration

8-4

The system can have a maximum of 32K of core memory. As addi-
tional fields of core memory are added, they permit overlapping the
running time of one user program with the swapping time of another,
expanding the resident Monitor to buffer a larger number of consoles
and reducing the amount of Monitor overlay required, thus increasing
operating speed.

With a minimum of 12K of core memory, the following options are
available.

1. Up to three DECdisks can be added to increase the number of
active users and their file storage.

2. Up to eight DECtapes drives can be added for individual users.

3. Memory Parity

4. Extended Arithmetic Element (PDP-8/I only)

5. Power Failure Protection

6. One High-Speed Paper Tape Punch

7. One DA-10 Interface (PDP-10)

8. One Bit Synchromous Communication Unit (Type 637)

9. One Line Frequency Clock (required with PT08's)

10. Maximum total of 16 active Teletype consoles with appropri-
ate Teletype Controls (PT08's or DC08)

The Data Communication System, Type DC08, is used in full duplex
mode. It consists of a Data Line Interface Unit (DL8/I), a Serial Line
Multiplexer Unit (685), and other devices connected to form a data
link and message switching system between the user consoles and the
central processor.

THE TSS/8 MONITOR

The TSS/8 Monitor is composed of the following routines.
Scheduler Error Message Handler

System Interpreter Buffer Handler

IOT Trap Handler DC 08 Service

Storage Allocator Disk Service

Overlay Control Optional Device Service

File Control

8-5

With the above routines, Monitor provides services which can be di-
vided into three broad categories : device service, scheduling and run-
ning, and communication.

On an interrupt basis (program interrupt — see Chapter 5) the device
service routines receive information from all input devices, distribute
that information out to the appropriate buffers, and inform the activa-
tion routine when a user program must be activated to receive its in-
formation. The device service routines also accept output from user
programs, buffer it, and send it to output devices whenever the devices
are able to receive it.

Scheduling is handled by a round-robin scheduling algorithm with
the exception that programs with disk requests pending are run out of
turn to optimize disk usage. The scheduling routine decides whether to
remove the current program from core, and if so, which program is to
be run next.

A user program is, at any point in time, in one of the following
states.

Running. The user program is in execution. It continues execution
until its quantum has expired or until it issues an I/O request that
cannot be satisfied immediately.

Active. The user program is ready to run and will be swapped into
core when its turn comes. A program can become active when

a. An output buffer is almost empty, thus assuring continuous
output of information,

b. An input buffer is almost full,

c. Input requested by a user has arrived,

d. A user determined activation condition has been satisfied, or

e. The user commands the system to begin execution.
Waiting. In this state the program would be active except that it is
waiting for the completion of some I/O request or special condition.
Dormant. The program is not being entered into the round-robin.
The user may be communicating with Monitor or the program may be
dismissed for a variety of Monitor-determined-reasons (e.g., illegal op
code).

A program may be swapped out of core memory for any of the fol-
lowing reasons.

1. The quantum of time has expired. The program moves from
the running state to the active state.

8-6

2. The program has requested that the quantum of time be ter-
minated. The program moves from the running state to the
active state.

3. The program has filled an output buffer, or has requested in-
put, but the input buffer is empty, or has requested a special
dismissal condition. The program moves from the running to
the waiting state.

4. The program has tried to execute an illegal instruction. The
program moves from the running state to the dormant state.

System Interpreter

Monitor checks all incoming characters for the call (CTRL/B) char-
acter. When the call character is encountered, Monitor routes all sub-
sequent characters up to and including the first carriage return (RE-
TURN key) to its System Interpreter's input buffer. Monitor's System
Interpreter performs the following services.

1. Verifies the user's account number and password when he logs
in, and provides him with a disk track to store user programs.
It releases facilities owned by the user when he logs out.

2. Parcels out extra consoles and input/output devices.

3. Provides commands for creating, opening, and maintaining the
user's files.

4. Allows the user to save all or part of his binary user program
for future use and reference, and restores it with its state un-
changed.

5. Provides accounting of console time and user program run
time.

6. Provides the user with information about the state of his pro-
gram while it is running, as well as information about the state
of the character control tables.

7. Provides commands for calling the various utility programs.

Communications to the System Interpreter are automatically du-
plexed. That is, all characters from a console keyboard to the System
Interpreter appear on that console's printer, regardless of the setting of
the character control tables (explained later).

The System Interpreter is kept quite busy receiving messages from
many keyboards and user programs, therefore, it must be run often.
Consequently, it occupies a priority position in the round-robin, being
scheduled whenever there are characters in its input buffer.

8-7

Phantom Routines

Monitor has two phantom routines: error and file control. They are
not phantoms as defined by Webster but they appear to act as though
they were. These routines reside on the disk, "jump" into core when
needed, perform their assigned tasks, and "fade" back onto the disk.

Phantoms are privileged routines which run in place of a regular
user program in the round-robin. The time the phantom takes to per-
form its service is charged to that user program upon which it is serv-
icing. For example, the error phantom prints all error messages for
running user programs, and printing a lengthy error message may re-
quire several intervals of time to which the program in error will be
charged. The error phantom routine replaces the offending program in
the round-robin and prints an error message, thus punishing the user
responsible for the error and no one else.

The file control phantom routine handles all modifications to the
user's files, such as creating, lengthening, renaming, and destroying
files. This phantom routine is brought into operation when a user pro-
gram executes certain input/output transfer (IOT) instructions or when
Monitor is acting for the user program.

Character and Data Flow

When a user logs into the system his account number and password
are associated with a job number, and that job number is then associ-
ated with a disk swapping track and the console(s) he owns. The ac-
count number establishes ownership of the input buffer attached to the
user program and the output buffer (s) attached to the printer (s) of
his console(s).

Monitor provides for communication among the users, user pro-
grams, and Monitor. Data flow is illustrated in Figure 8-2. Communi-
cation is provided through the IOT instructions. When a program exe-
cutes IOT instructions, Monitor picks up locations in the user program
as parameters to service routines. These routines may simulate I/O
to or from the online device, control or release ownership of devices,
handle character transmission, or return information to the user pro-
gram by filling core locations within the program. Through IOT in-
structions the user program can make its wants known and Monitor
can inform the user program of any unusual conditions in the system.

Characters are generated by users typing at keyboards or by user
programs typing out. Characters go into input buffers to be read by
user programs or to output buffers to be printed on console printers.
The character control tables (Figure 8-2) set up useful character trans-
mission paths, whereby any character source (keyboard and program)

8-8

CHARACTER
CONTROL

REQUEST

DISK \

tracks)
t swapping

SERVICE OR I/O

USER
PROGRAM
4K COKE

SWITCHBOARD

KEYBOARD

I/O
CONTROL

1 ,

-J
o

GENERAL

INPUT
ROUTINE

INPUT

BUFFER

■a —

m <->

PRINTER

OUTPUT

GENERAL

INPUT
ROUTINE

BUFFER

PERMISSION

1 '

COMMAND
LANGUAGE

AND
SERVICE

T .

OTHER
I/O

REQUEST
FOR SPECIAL
SERVICE
OR I/O

PHANTOM
SERVICES

DISK
FILES

Figure 8-2. Character and Data Flow

may, with permission, send to any character buffer. To insure that
unwanted connections are not made, using the appropriate command
to Monitor the user owning any character buffer may grant or deny
permission for connection into that buffer.

The character control tables are a permission table and a switch-
board table. The permission table allows a user to control which char-
acter sources may place characters in which of his buffers, and the
switchboard table controls the actual routing of these characters. Char-
acters originating from any keyboard or from any user program typing
out can, with permission, be placed in any or all of the keyboard in-
put buffers or any or all of the printer output buffers. The user, using
appropriate Monitor commands, establishes the actual routing of char-
acters to and from his associated buffers.

The character control feature finds applications in:

1. Duplexing — Characters typed at a keyboard can be printed
on other consoles without user program intervention.

2. Interprogram Communication — User programs can communi-
cate with each other and with Monitor. Thus, several user
programs may run as one system coordinating their separate
tasks through character communication.

8-9

3. Interconsole Communication — Users at consoles can set up
general links for conferences, teaching, monitoring, or for any
number of reasons.

4. Multiple Consoles — User programs may receive characters
from and send characters to more than one console. This al-
lows a user program to act as a time-sharing subsystem within
TSS/8, controlling its own set of consoles, e.g., as in teaching
machine monitors.

MONITOR COMMANDS

A Monitor command is a string of characters terminated by a semi-
colon (;) or a carriage return (RETURN key). Commands to Monitor
are typed on the console keyboard by the user or output by the user
program, with each command beginning with a command name. In
some cases, the command name is the entire command, in which
case it is followed directly by a terminator. In other cases, the com-
mand name is followed by a space, one or more parameters, and then
a terminator (see example below).

.TIME (Requests the time of day. This command was termi-

nated by the RETURN key)

.TIME d (Requests the processing time used by job Ci. Termi-

nated by the RETURN key)

Only enough characters need be typed in the command name to
uniquely identify the command name as shown in the following ex-
ample.

LOGI for LOGIN

LOGO for LOGOUT

More than one command may be typed on a line, with all but the
last command being terminated by a semicolon; the last command is
terminated by the RETURN key (see example below). Commands
are executed only when the RETURN key is typed, which explains why
the last command on a line must be terminated by the RETURN key.

.OPEN c, Sa c 3 ; LOAD d &,; START c x

As shown in the above examples, each command or a line of com-
mands is typed after the dot typed by Monitor. The dot is typed by
Monitor when it is ready and available to accept commands from the
user.

8-10

Parameters may be typed as octal numbers, decimal numbers, char-
acter strings, or single letters. In the following descriptions of Monitor
commands the parameters are coded as follows.

d, c„ . . . represent octal numbers

d t , d 2 , . . . represent decimal numbers

Si, Sj, . . . represent character strings

li, U, . . . represent single letters

Logging In and Out

Logging in and out of TSS/8 is a function performed by Monitor's
System Interpreter routine. When a console is in the free state, a pro-
spective user may attempt to log into the system,

This is a request by a user to enter into the system. If the console
from which the command is typed is free, there is an available disk
track, and the two parameters following the command LOGIN form
a legitimate account number and password, then the user will be
logged into the system.

Ci is the user's account number.
s a is the user's password.

At the time of login, the switchboard table is initialized to the nor-
mal operation setting. However, Monitor diverts all characters typed
to the System Interpreter until the user gives a command that indi-
cates otherwise.

LOGOUT;

This is a request by a user to leave the system. It disconnects all con-
, soles and temporary disk track that he owns, places his programs
in the free state, and resets the switchboard table. It also writes an
account record on the disk showing how much computing time
(processing time) and console time was used by the user's pro-
gram(s).

TIME &;

This is a request for Monitor to type out the computing (processing)
time used since login. If job Ci is not specified, the job owning the
console is assumed. If requested before login and if no job is spe-
cified, the time-of-day is typed. If, at any time, job is specified
the time-of-day is typed.

8-11

Console Manipulation

ASSIGN K c;

This is a request for console number Ci to be added to those consoles
already owned by the requesting user.

K denotes console

Ci is the console number

If the console is free, it will be given to the user and the user's
switchboard table settings will be augmented by those additional
entries which put the new console in the normal state with respect
to its own printer and the program on the user's disk track. Cross
settings for interconsole connections are left to the user's discretion.

SLAVE c;

This command causes console number Ci to be slaved. This means
that console number c, will be unable to communicate with Monitor,
that the system will ignore the call (CTRL/B) character from that
console. However, the console may serve as an input/output device
for the user's program. Monitor checks to make sure that the re-
questing user owns the console he wishes to slave. It is illegal to
slave every console that a user owns.

UNSLAVE Cu

This command restores a slaved console to normal status. After this
command, console number Ci will be able to communicate with
Monitor. The user must own the console to unslave it.

RELEASE K c;

This command releases (deassigns) console number Ci and puts it
in the free state. Monitor checks to make sure that the requesting
user owns the console he wishes to release. When the console is
released, the switchboard table is reset and the input and output
buffers are cleared.

Device Allocation

ASSIGN I,;

This is a request for access to the device specified by L
\ — R for high - speed paper tape reader

P for high - speed paper tape punch

C for card reader

L for line printer

I for incremental plotter
If device K is not busy, it will be allocated to the user program issu-
ing the command. If device h is busy, the job number of the user
having possession is returned. A RELEASE h; or LOGOUT; will re-
lease the device.

8-12

ASSIGN D &;

This command is the same as ASSIGN L;
D denotes DECtape unit
Ci is the unit number of the DECtape unit
RELEASE L;
RELEASE K &;
RELEASE D Cl ;
Each of these commands will annul the current assignment of the
specified device.

File Control

OPEN d Si c 2 ;

This command establishes association between an internal file num-
ber and a file. After this command is given, the file of account c,>
with the name s t is associated with internal file number c,'. The in-
ternal file number specified must be between zero and three inclusive.
If c 2 is not specified, the account number of the current user is as-
sumed.

CLOSE & Cs . . . ;
This command closes the files specified by d c. . . ,

Ci c 2 . . . is a list of internal-file numbers separated by spaces
After this command is given, no writing can be done on the files
specified, and the associations between the internal file numbers and
the files are broken.

CREATE &; ■

This command causes Monitor to create a new file which is to have
the name s u if there is available space in the File Storage Area of
disk. At the time of creation, Monitor will enter the name of the
new file and the date of creation into the owner's file directory.
Example:

CREATE NEWF; Asks Monitor to create a new file named
NEWF. If there is no more space in File Stor-
age, Monitor will so inform the user.
RENAME Cl Sl ;

This command renames a file. The file to be renamed must be al-
ready open and associated with internal file number c u Its new name
will become &. Example:

RENAME 1 MSTR; Assume that file NEWF has been opened to
internal file number 1; this command will
change the name of that file from NEWF to
MSTR.

8-13

REDUCE c, d,;

This command reduces the length of a file. The file which is to be
shortened must be open and associated with internal file number &.

d t is the number of segments to be removed from the end of the file

If di is greater than or equal to the number of segments in the file,
the file is deleted from the directory. Example:

REDUCE 2 2; Assume file TT13 is opened to internal file

number 2; this command then removes 2 seg-
ments from its end and returns those segments
to free storage.

EXTEND c, d,;

This command extends the length of a file. The file which is to be
lengthened must be open and associated with internal file number c t .

d x is the number of segments to be added to the end of the file

If there is free storage space available, Monitor will lengthen the file
as requested.
PROTECT c, c»;

This command changes the file protection mask of a file. The file
to be protected must be open and associated with internal file num-
ber Ci.

c 2 is the new file protection mask

For file protection, the 12-bit account number is partitioned into a
project number (high order 7 bits) and a programmer number (low
order 5 bits).
File protection masks (c 2 ) are assigned as follows:

Co = 1 read protect against users whose project number differs from

owner's.
e, = 2 write protect against users whose project numbed differs from

owner's.
c 2 = 4 read protect against users whose project number is same as

owner's.
c 2 = 10 write protect against users whose project number is same as

owner's.
c a = 20 write protect against owner.
c = can be the sum of any of the above values.

The protect command is illegal for all users except the owner.
Example:

PROTECT 1 3; Read and write protect internal file 1 against

access by any user whose project number dif-
fers from the owner's.
8-14

F &;

This command causes Monitor to print out the current state of the
association of the user's internal file number Ci with the file. The
response format is c= s t c 3 &
where

c a is the owner's account number.

Sj is the filename.

c 3 is the protection mask.

dj is the number of segments.

Control of User Programs

START &;

This command begins execution of a user program at location &. In
addition, the command resets the switchboard table so that all
characters typed from either a keyboard or program directed into
the user program's input buffer are no longer intercepted by the
System Interpreter.

START;

This command restarts a user program. If Monitor has been called
during the execution of the user program, the complete state of the
program is saved including the location of the next instruction to be
executed. When the START command is given, the program's state
is restored and the program continues execution where it left off. As
in the START &; command, characters intended for the user pro-
gram's input buffer are no longer intercepted by the System Inter-
preter.

DEPOSIT d c». ,.c n ;

This command stores c 2 in location c,; c 3 in location & +1; . . .;c n
in location & +n — 1.
n is equal to or less than 10 (decimal) .

A user can load a binary user program using the DEPOSIT com-
mand, although it is. much easier using DDT-8. This command is
useful when making small patches to stored programs. The address
Ci must be an absolute address within the user's 4K core area.
EXAMINE d &;,

This command causes Monitor to type the contents of the di loca-
tions starting at location &.

di is equal to or less than 10 (decimal).

If di is not specified, the contents of location & is typed. The address
Ci must be an absolute address within the user's 4K core area.

8-15

Saving and Restoring Binary User Programs. The SAVE, LOAD, R,
and RUN commands leave file s x open and assign it internal file num-
ber 3, and turn the user's interrupt system off.
SAVE d s»
SAVE d s t d;
SAVE Ct s,. c. c 3 ;
SAVE d Si c 2 Cs d;
These commands write portions of the user's core image onto a file
whose owner's account number is &, filename is Si (Si must have
been previously CREATEd), and

d is the file address of the first word to be written; if not specified,

the entire 4K is written on the first 4096 words of the file.
d is the core address of the first word to be written; if not specified,

all 4096 words are written.
d is the core address of the last word to be written; if not specified,
7777 is assumed.
If Ct is not specified, the account number of the current user is
assumed.
Example:

SAVE NEWF; Writes core words through 7777 or words

LOAD Ci s,- through 7777 of file NEWF.

LOAD Ci Si c,.;

LOAD Ci Si Cj. C:>;

LOAD Ct Si d d cr,

These commands read certain portions of the file whose owner's
account number is c, and whose filename is St, into core. Ct may be
omitted if it is the same as the account number under which the
user is logged in.

c, is the file address of the first word to be read; if not specified,
words through 7777 are read into words through 7777.

d is the core address of the first word to be loaded; if not specified,
all words through 7777 are loaded.

c, is the core address of the last word to be loaded; if not specified,
7777 is assumed.
Example:

LOAD NEWF 5 10 17; Loads words 5 through 14 into words 10

„ through 17 respectively.

R Si;

This command is equivalent to

OPEN 3 s, 2; LOAD 2 s,; START

which loads program Si from the system library (account 2) and
starts the program running.

8-16

RUNs.;

This command is equivalent to

OPEN 3 s,; LOAD s^ START
RUN ci Sx;

This command is equivalent to

OPEN 3 Si d; LOAD c t s x ; START

This command stops the execution of the user program, saves its
complete state, and sets the switchboard table so that all characters
directed to the program's input buffer will be intercepted by System
Interpreter.

WHERE;

This command causes Monitor to type out the current state of a
user program's accumulator, link, program counter, and switch
register, multiplier quotient, and step counter.

USER;

This command causes Monitor to type out the number of the job
and devices owned by the user.

USER c;

This command causes Monitor to type out the numbers of the de-
vices owned by user d. If job is specified, the numbers of unas-
signed devices are typed.

SWITCH &;

This command sets the user's switch register to Ci.

BREAK a;

This command sets the user's keyboard delimiter mask to the value

Ci.

Permission and Switchboard Tables

Using Monitor commands the user may set, reset, and read any
given bit in either the permission table or the switchboard table.
SET L h d 1 3 c 2

This command sets a bit in either the permission or switchboard
table to a 1.

li = P denotes permission table

S denotes switchboard table
I»= K denotes keyboard

P denotes user program
Ci is the octal number of the keyboard or program
1 3 = I denotes input buffer

O denotes output buffer
c a is the octal number of the buffer which is to receive the charac-
ters from the character source

8-17

Monitor checks for appropriate ownership to decide whether it will
allow the connection to be made. The ownership of the character
source is used to determine legality in setting the switchboard table
while the ownership of the character sink is used in setting the per-
mission table. Examples :

SET P K 13 O 12; Requests that keyboard number 13 be given
permission to write on the printer (output buf-
fer) of console 12. .

SET S P 4 I 16; Informs Mnoitor that job number 4 would

like to type characters into the input buffer of
user program 16. This request will be granted
only if user 1 6 has previously given permission
by SET P P 4 I 16; or by the equivalent IOT
instruction (SSP).

RESET Ii Lake,;

This command is identical in all respects to the SET command above
except that the bit indicated by the parameters will be cleared to
zero.

READ h 1. c, L a;

This command uses the same parameters to specify a bit in either
the permission or switchboard table as in the SET and RESET
commands above. Monitor will type out the value of the bit.

DUPLEX;

This command is a shorthand command to set the switchboard table
so that characters typed on the keyboard from which this command
is issued will appear on the printer of that console.
Example:

DUPLEX; If the keyboard issuing this command is num-

ber 16, the command is equivalent to RESET
SK16 16;
UNDUFLEX;

This command is a shorthand command to undo the effects of the

DUPLEX command.

ALLOW a;

This command is shorthand to indicate that a user is giving permis-
sion for keyboard number a to place characters in the output buffer
of his console.
Example: ,

ALLOW 4; Assume that this command was issued from

keyboard number 7, then it is equivalent to
SET PK407;
LINK a;

This command is given by a user who wishes to communicate with
a console other than the one at which he is sitting. This command

8-18

is legal only if the owner of that console has set the permission table
so that he will accept characters from the requesting console into
his printer's output buffer. The command is shorthand for the com-
mand ALLOW Ci; followed by the command to set the switchboard
table so thaf characters from the requesting console will be placed
in output buffer number Ci. Example:

LINK 16; Assume that the console which issued this com-

mand was number 7, then this command would
be legal only if the owner of console number
16 had previously set the permission table to
allow keyboard 7 access to that output buffer.
He could have done this with ALLOW 7; or
by SET P K O 16; or by an equivalent IOT
instruction executed by his program. If that;
permission has been granted, the LINK com-
mand is equivalent to the following two switch-
board commands: SET P K 16 O 7; (AL-
LOW 16) and SET S K 7 O 16;

Inter-System Communication

In those TSS/8 systems having a local connection to a PDP-10
Time-Sharing System or a Synchronous Data Communication System
(Type 637), the input to the user's input buffer and program output is
scanned for the character sequences CTRL/B CTRL/X and CTRL/B
CTRL/Y, respectively. All characters up to the next CTRL/B are
diverted to the PDP-10 or 637 System, whichever the case may be.
Characters from the PDP-10 and 637 System are directed into the
user's input buffer.

INPUT/ OUTPUT TRANSFER INSTRUCTIONS

Whenever a user program executes an input/output transfer (IOT)
instruction (an instruction of the form 6XXX) the system traps the
instruction and transfers control to a system service or simulation rou-
tine. These routines accomplish special tasks, set up parameters for
the system, and perform input/output for the user program.

IOT instructions may be separated by function into three types:

1. Input/ output instructions available on the standard PDP-8/I
without a Time-Sharing Monitor— -When a user program exe-
cutes one of these IOTs, TSS/8 simulates an input/output
function similar to the function of the IOT instruction on the
standard PDP-8/I. Some standard PDP-8/I IOT instructions
are illegal in TSS/8 (see Appendix C of Time-Sharing System
—TSS/8 Monitor, Order No. DEC-T8-MRFB-D) .

8-19

2. IOT instructions to request input/output service from TSS/8
unavailable on the standard PDP-8/I — These include requests
for DECdisk, DECtape, high-speed paper tape reader and
punch, card reader, and console character handling.

3. IOT instructions which call subroutines to set user param-
eters or to alter the time-sharing environment for a particular
user program — An alteration may, for example, include re-
quests to add or release facilities or to change mode of
character handling.

An IOT instruction usually acts as a subroutine call. Therefore, de-
pending on the specific IOT instruction, parameters may be loaded
into the accumulator (AC) before execution of the IOT. In some cases,
the parameter in the AC acts as a pointer to a parameter block in the
user's program. If the system has to return information to the user's
program, it returns that information in the AC or in a block of loca-
tions in the user's program (the beginning address of this block is in
the AC).

In general, nearly all TSS/8 console commands have corresponding
IOTs which allow the user the same features under program control
which he has with the console commands.

ERROR MESSAGES

User Program

Error messages are typed on the user's printer by the error phantom
routine when error conditions occur in a running user program. If the
user program is not enabled for system error interrupts, the error mes-
sages are typed in the following format.

s, FOR USER d.

AC = c„, L = c 3 , PC = c,

INSTR = c
Si is a string describing the nature of the illegal instruction c 5 that user
program c. has executed at location c At the time the illegal instruc-
tion was executed the value of the accumulator was c» and the value
of the link was c>. For example, an error message might appear as fol-
lows:

8-20

ILLEGAL IOT FOR USER 2
AC = 1520, L = 1, PC = 3552
INSTR = 6025

The user program has executed an
IOT which the system regards as
illegal. The illegality may be for
one of two reasons:
1. The IOT itself may be illegal.
2. The parameters of a legal IOT
may invalidate it.

System Interpreter

The following error messages result from illegal requests to the Sys-
tem Interpreter. They are printed by the error phantom on the console.

'fc 1

ILLEGAL REQUEST

SWITCHBOARD ERROR

CONSOLE IN USE
LOGOUT TO RELEASE

UNAUTHORIZED
ACCOUNT

FULL

Cx HAS IT
FAILED BY d x

The System Interpreter does not under-
stand the command.
Si = command
The user has requested an illegal service.
This error usually results when some pa-
rameter has been given an incorrect value
or the request refers to a facility not
owned by the user.

The user has attempted to set a bit in the
permission table not owned by him; or
the user has attempted to make a connec-
tion in the switchboard table for which
permission has not been granted.

The user has tried to log in on a console
which is already in use; or he has at-
tempted to add a console to those he
already owns.

The user has attempted to use a console
which is not logged in.

The user has attempted to log into the
system with an invalid account number
or name.

This message may appear on an attempt
to log into TSS/8 from a console. It
means that all disk swapping tracks are
in use.

Job number c x has the request device; the
request for its acquisition cannot be
granted.

The user has attempted to extend a file
beyond the available disk storage, di seg-
ments of the user's request could not be
added.

8-21

FILE NOT OPEN

DIRECTORY FULL

PROTECTION
VIOLATION

FILE NOT FOUND

DISK FULL

The user has attempted to manipulate a
file which is not associated with an in-
ternal file number.

The user's file directory has no room for
new file names.

The user has attempted to use a file in a
manner contrary to the file protection
code specified for that file.

The system could not find the desired file
in the specified directory.

The user had attempted to create a new
file when there is no room on the disk
for it.

CONCLUSION

The above is merely an overview of TSS/8. There are many other
aspects not mentioned here. For a thorough coverage, see Time-Shar-
ing System— TSS/8 Monitor, Order No. DEC-T8-MRFB-D.

8-22

Chapter 9

FOCAL
Programming

FOCAL (FOrmula CALculator) is an online, conversational, inter-
pretive language for the PDP-8 family of computers. It is designed to
help students, engineers, and scientists solve numerical problems. The
language consists of short, easy-to-learn, imperative English statements.
Mathematical expressions are typed in standard notation. The best way
to learn the FOCAL language is to sit at the Teletype console and try
the commands, starting with the examples given in this chapter.

FOCAL puts the full calculating power and speed of the computer
at your fingertips. FOCAL is an easy way of simulating mathematical
models, plotting curves, handling sets of simultaneous equations in
n-dimensional arrays, and much more. A few of the many kinds of
problems that have been solved by FOCAL are described under "Ex-
amples of FOCAL Programs." The user can become acquainted with
many applications of FOCAL by duplicating the example programs
using different variables.

This chapter describes the features of FOCAL, 8/68, which is issued
on tape DEC-08-AJAC-PB.

EQUIPMENT REQUIREMENTS

FOCAL operates on a 4K PDP-8 family computer with an ASR 33
Teletype, and with or without a high-speed reader/punch, analog-to-
digital converter (189), oscilloscope display (34D), or any other DEC
peripherals with the appropriate program overlays available from DEC.

GETTING ONLINE WITH FOCAL

The FOCAL program is furnished to the user on punched paper
tape in binary-coded format. Therefore, it is loaded into core using the
Binary Loader program, as described in Chapter 6.

9-1

The Initial Dialogue

After FOCAL has been loaded and started, it begins typing out its
initial dialogue, giving the user the options of retaining certain groups
of mathematical functions. If these functions are not needed, the user
answers FOCAL's questions by typing NO and the RETURN key, and
FOCAL erases those functions from core, thus the user gains additional
core storage for use by his programs.

Samples of the initial dialogue are shown below.

CONGRATULATIONS!!

YOU HAVE SUCCESSFULLY LOADED 'FOCAL' ON A PDP-8/1 COMPUTER.

SHALL I RETAIN LOG, EXP, ATN ?:YES
PROCEED.

With the above response, all mathematical functions are retained,
and the user has about 720 locations available for his programs.

CONGRATULATIONS!!

YOU HAVE SUCCESSFULLY LOADED 'FOCAL' ON A PDP-8/1 COMPUTER.

SHALL I RETAIN LOG. EXP, ATN ?:NO
SHALL I RETAIN SINE, COSINE ? :YES

PROCEED.

When the user answers NO to the first question FOCAL asks a
second question. The above response leaves about 975 locations avail-
able for the user's programs.

CONGR ATU LATIONS! !

YOU HAVE SUCCESSFULLY LOADED 'FOCAL' ON A PDP-8/1 COMPUTER.

SHALL I RETAIN LOG, EXP, ATN ?:NO
SHALL I RETAIN SINE, COSINE ? :NO

PROCEED.

The above response erases all mathematical functions from core,
giving t the user about 1105 locations for use by his programs.

A simple FOCAL program which determines the number of core
locations available for the user's programs and a formula for calculating
the length of a user program are given under "Estimating the Length
of a User's Program."

In the second line of the initial dialogue, FOCAL identifies the type
of computer being used— PDP-8/1, PDP-8/L, PDP-8, etc. FOCAL
concludes the initial dialogue by telling the user to PROCEED, fol-
lowed by an *, and waits for user input.

9-2

THE FOCAL LANGUAGE

When the initial dialogue is concluded, FOCAL types
*

indicating that the program is ready to accept commands from the user.
Each time the user completes typing a Teletype line and terminates it by
depressing the RETURN key or after FOCAL has performed a com-
mand, an asterisk is typed to tell the user that FOCAL is ready for
another command.

Simple Commands

One of the most useful commands in the FOCAL language is TYPE.
To FOCAL this means "type out the result of the following expres-
sion." When you type (following the asterisk which FOCAL typed),

♦TYPE 6.4318+8.1346

and then press the RETURN key, FOCAL types

= 14.5664*

Another useful command is SET, which tells FOCAL "store this
symbol and its numerical value. When I use this symbol in an expres-
sion, insert the numerical value." Thus, the user may type,

*SET A=3. 14159; SET B=428.77; SET C=2.71828

*
The user may now use these symbols to identify the values denned in .
the SET command. Symbols may consist of one or two alphanumeric
characters. The first character must be a letter, but must not be the
letter F.

♦TYPEA+B+C

= 434.6300*

Both the TYPE and SET commands will be explained more fully in
their respective sections of this chapter.

FOCAL is always checking user input for invalid commands, illegal
formats, and many other kinds of errors, and types an error message
indicating the type of error detected. In the example,

*HELP

703.31
*TYPE 2++4

?07.<0

HELP is not a valid command and two plus signs (double operators)

9-3

are illegal. The complete list of error messages and their meanings is
given under "Error Diagnostics."

Output Format

FOCAL is originally set to produce results showing up to eight digits,
four to the left of the decimal point (the integer part) and four to the
right of the decimal point (the fractional part) . Leading zeros are sup-
pressed, and spaces are shown instead. Trailing zeros are included in
the output, as shown in the examples below.

*SETA=77.77;SETB=1111.1111; SET C=39
•TYPE A,B,C

= 77.7700= 1111.1100= 39.0000*
The results are calculated to six significant digits. Even though a re-
sult may show more than six digits, only six are significant, as shown
above in SET B = 1111.1111, which FOCAL typed as = 1111.1100.
The output format may be changed if the user types

TYPE %x.yz,

where x is the total number of digits to be output and yz (always two
digits, i.e., 01, 08, 12, etc.) is the number of digits to the right of the
decimal point, x and yz are positive integers, and the value of x cannot
exceed 19. When first loaded, FOCAL is set to produce output having
eight digits, with four of these to the right of the decimal point
(%8.04). For example, if the desired output format is mm.nn, the
user may type

*TYPE %4.02, 12.22+2.37
and FOCAL will type

= 14.59*
Notice that the format operator (%x.yz,) must be followed by a
comma.

In the following examples, the number 67823 is typed out in several
different formats.

*SET A=67823
*TYPE%6.01,A
= 67823.0*
*TYPE%5,A
= 67823*
*TYPE%8.03, A
= 67823.000*

9-4

If the specified output format is too small to contain the number,
FOCAL automatically prints the number in floating-point format, as
explained below. If the specified format is larger than the number,
FOCAL inserts leading spaces:

*TYPE%7, 67823
= 67823*

Leading blanks and zeros in integers are always ignored by FOCAL.

TYPE % 8.04, 0016, 0.016, ., 007

= 16.0000= 0.0160= 0.0000= 7.0000*

Floating-Point Format

To handle much larger and much smaller numbers, the user may re-
quest output in exponential form, which is called floating-point or E
format. This notation is frequently used in scientific computations, and
is the format in which FOCAL performs its internal computations. The
user requests floating-point format by including a % followed by a
comma, in a TYPE command. From that point on, until the user again
changes the output format, results will be typed out in floating-point
format.

♦TYPE %, 11

= 0.110000E+02*

This is interpreted as .11 times 10 2 , or simply 11. Exponents can be
used to ±616. The largest number that FOCAL can handle is
+0.999999 times 10« 16 , and the smallest is —0.999999 times 10 61B .
To demonstrate FOCAL's power to compute large numbers, you can
find the value of 300 factorial by typing the following commands. (The
FOR statement, which will be explained later, is used to set I equal
to each integer from 1 to 300.)

*SETA=1

*FOR 1=1,300; SET A=A*I .

*TYPE % , A ( wa it for FOCAL to

= 0.306051E+615* , calculate the value)

Arithmetic Operations

FOCAL performs the usual arithmetic operations of addition, sub-
traction, multiplication, division, and exponentiation. These are written
by using the following symbols:

9-5

Symbol Math Notation FOCAL

tExponentiation

3 3

3f3

(Power must be a
positive integer)

*Multiplication

3-3

3*3

/Division

3-^3

3/3

+Addition

3+3

—Subtraction

3-3

These operations may be combined into expressions. When FOCAL
evaluates an expression, which may include several arithmetic opera-
tions, the order of precedence is the same as that in the list above. That
is, exponentiation is done first, followed by multiplication, division,
addition and subtraction. Addition and subtraction have equal priority.
Expressions with the same precedence are evaluated from the left to
right.

A+B*C+DisA+(B*C)+Dnot(A+B)*(C+D) nor (A+B)*C+D
A*B+C*Dis (A*B)+(C*D) not A*(B+C) *D nor (A*B+C)*D

X/2*Yis —
2Y

2T2T3 is 4 3 not 2 s

Expressions are combinations of arithmetic operations or functions
which may be reduced by FOCAL to a single number. Expressions
may be enclosed in properly paired parentheses, square brackets, and
angle brackets (use the enclosures of your choice for clarity; FOCAL
is impartial and treats them all in the same way).
For example,

*SETA1=(A+B)*<C+D>*[E+G]

* -

The [ and ] enclosures are typed using SHIFT/K and SHIFT/M, re-
spectively.

Expressions may be nested. FOCAL computes the value of nested
expressions by doing the innermost first and then working outward.

*TYPE%, [2+(3-<l*l>+5)+2]
= 0.110000E+02*

Note that the result is typed in floating-point format.

More About Symbols

The value of a symbolic name or identifier is not changed until the
expression to the right of the equal sign is evaluated by FOCAL. Tfiere-

9-6

fore, the value of a symbolic name or identifier can be changed by re-
defining it in terms of itself (i.e., in terms of its current value).

*SETAl=3f2

*SETA1=A1 + 1

♦TYPE % 2, Al

= 10*

NOTE

Symbolic names or identifiers must not begin with the letter F.

The user may request FOCAL to type out the values of all of the
user-defined identifiers, in the order of definition, by typing a dollar

sign($).

TYPE % 6. 05, $
The user's symbol table is typed out like this

A@(00)= 0.30605 1E+6 1 5
B@(00)= 1111.11
C@(00)= 39.0000
I@(00)= 301.000
A1(00)= 10.0000
D@(00)= 0.00000
E@(00)= 0.00000

G@(00)= 0.00000

If an identifier consists of only one letter, an @ is inserted as a second
character in the symbol table printout, as shown in the example above.
An identifier may be longer than two characters, but only the first two
will be recognized by FOCAL and thus stored in the symbol table.

Subscripted Variables

FOCAL always allows identifiers, or variable symbols, to be further
identified by subscripts (range ±2047) which arc enclosed in paren-
theses immediately following the identifier. A subscript may also be an
expression:

*SETA1(I+3*J)=2.71;SETX1(5+3*J)=2.79
The ability of FOCAL to compute subscripts is especially useful m
generating arrays for complex programming problems. A convenient
way to generate linear subscripts is shown under "Simultaneous Equa-
tions and Matrices."

The ERASE Command

It is useful at times to delete all of the symbolic names which you
have defined in the symbol table. This is done by typing a single com-

9-7

mand: ERASE. Since FOCAL does not clear the user's symbol table
area in core memory when it is first loaded, it is good programming
practice to type an ERASE command before defining any symbols.

Handling Text Output

Text strings are enclosed in quotation marks (". . .") and may in-
clude most Teletype printing characters and spaces. The carriage return,
line feed, and leader-trailer characters are not allowed in text strings.
In order to have FOCAL type an automatic carriage return/line feed at
the end of a text string, the user inserts an exclamation mark(!).

*TYPE "ALPHA"!"BETA"!"DELTA"!

ALPHA

BETA

DELTA

To get a carriage return without a line feed at the end of a text type-
out, the user inserts a number sign (#) as shown below.

*TYPE !"

X Y

Z"#'

+ ="#"

■ 4 SPACES

■ 1 SPACE

• 3 SPACES
. 5 SPACES

■ 2 SPACES
. 8 SPACES

/"!

X+Y # Z

The number sign operator is useful in formatting output and in plotting
another variable along the same coordinate (an example is given under
"Intercept and Plot of Two Functions") .

Indirect Commands

Up to this point we have discussed commands which are executed
immediately by FOCAL. Next we shall see how indirect commands are
written.

If a Teletype line is prefixed by a line number, that line is not exe-
cuted immediately, instead, it is stored by FOCAL for later execution,
usually as part of a sequence of commands. Line numbers must be in
the range 1.01 to 15.99. The numbers 1.00, 2.00 etc., are illegal fine
numbers; they are used to indicate an entire group of lines. The number
to the left of the point is called the group number; the number to the
right is called the step number. For example,

9-8

♦ERASE

»*1.1 SETA=1
*1.3 SETB=2
•1.5 TYPE%1,A+B
*

Indirect commands are executed by typing the GOTO or DO com-
mands.

The GOTO command causes FOCAL to start the program by exe-
cuting the command at a specified line number. If the user types

*GOT0 1.3

= 2*
FOCAL started executing the program at the second command in the
example above, so that the variable "A" was not previously defined and
therefore has a value of zero.

The GO command causes FOCAL to go to the lowest numbered
line to begin the program. If the user types a direct GO command after
the indirect commands above, FOCAL will start executing at line l.L

♦GO

= 3*

The DO command is used to transfer control to a specified step, or
group of steps, and then return automatically to the command immedi-
ately following the DO command.

♦ERASE ALL
*1.1 SETA=1;SETB=2
♦1.2 TYPE " STARTING "
♦1.3 DO 3.2

♦2.1 TYPE " FINISHED "
♦3.1 SETA=3;SETB=4
♦3.2 TYPE%1,A+B
♦GO
STARTING = 3 FINISHED = 7*

When the DO command at line 1 .3 was reached, the command TYPE
% 1, A+B was performed and then the program returned to line 2.1.
The DO command can also cause FOCAL to jump to a group of
commands and then return automatically to the normal sequence, as
shown in the example below.

♦ERASE ALL

♦1.1 TYPE "A "

♦1.2 TYPE"B "

♦1.3 TYPE"C "

♦1.4 DO 5.0

♦1.5 TYPE" END"; GOTO 6.1

9-9

*5.1 TYPE"D "

*5.2 TYPE "E "

*5.3 TYPE"F "

*6.1 TYPE". "

*GO

A B C D E F END. *

When the DO command at line 1.4 was reached, FOCAL executed
lines 5.1, 5.2, and 5.3 and then returned to line 1.5.

An indirect command can be inserted in a program by using the
proper sequential line number. For example,

*ERASE ALL

*4.8SETA=1;SETB=2

*6.3 TYPE % 5.4, B/C+A

*4.9SETC=1.31*.29

*GO

= 6.2645*

where line 4.9 will be executed before line 6.3 and after line 4.8.
FOCAL arranges and executes indirect commands in numerical se-
quence by line number, starting with the smallest line number and going
to the largest.

Error Detection

During execution, FOCAL checks for a variety of errors. When
an error is detected FOCAL stops execution, types a ? followed by an
error message, types an *, and waits for more user input. When the
error occurs in a direct statement, FOCAL types the error message
immediately after the user terminates that line (direct statements are
executed immediately after the line terminator). For example,

*SET A=2; PET B=4; TYPE A + B

703.31

PET is not a FOCAL command. Therefore, FOCAL issued the error
code 703.31 which means that an illegal command was used. (See
"Error Diagnostics" for a list of all error messages and their meanings.)
When an error occurs in an indirect statement the error message is
typed when FOCAL encounters that statement during execution. And
in addition to the error code FOCAL types the line number of the line
containing the error. For example,

•1.10 SET A=2; TYPE "A", A, !

*1.20 SET B=4; TYPE "B", B, !

*1.30 GO TO 1.10

♦1.40 TYPE "A+B", A+B

•GO

9-10

A= 2.0000
B= 4.0000

701.89 @ 01.30

FOCAL executed lines 1.10 and 1.20 and then recognized that GO
" TO is a 1-word command and should have been written GOTO. There-
fore, it issued the error message, meaning GOTO was not used as one
word at line number 1.30.

Corrections

If the user types the wrong character, or several wrong characters,
he can use the RUBOUT key, which echoes a backslash (--.) for each
RUBOUT typed, to erase one character to the left each time the RUB-
OUT key is depressed. For example,

♦ERASE ALL

*1.1 PYTYPEX-Y

*1.2 SET$=13 : \ \\ \X=13

♦WRITE

C— FOCAL, 1968

01.10 TYPEX-Y.

01.20 SETX=13
*

The left arrow ( <-) erases everything which appears to its left on
the same line, except when being used to correct a value typed after a
colon (:) in response to an ASK command (see "ASK").

*1.3TYPEA,B, C^-

*WRITE

C— FOCAL, 1968

01.10 TYPEX-Y

01.20 SETX=13
*

A line can be overwritten by repeating the same line number and
typing the new command.

♦14.99 SET C9(N+3)=15

is replaced by typing

♦14.99 TYPE C9/Z5-2
♦WRITE 14.99

14.99 TYPE C9/Z5-2

9-11

A line or group of lines may be deleted by using the ERASE com-
mand with an argument. For example, to delete line 2,21, the user
types.

*ERASE 2.21

To delete all of the lines in group 2, the user types

*ERASE 2.0

The user's entire symbol table is erased from memory whenever a
line number is retyped or the ERASE command is given. Since
FOCAL does not zero memory when loaded, it is good practice to
ERASE before denning symbols. The command ERASE ALL erases
all user input, i.e., program text and variables. Therefore, the ERASE
ALL command should be given before writing a new program.

The MODIFY command is another valuable feature, especially in
editing. It may be used to change any number of characters in a par-
ticular line, as explained under "MODIFY."

Alphanumeric Numbers (Using Letters as Numbers)

Numbers must start with a numeral but may contain letters. FOCAL
interprets as a number any character string beginning with a numeral,
through 9. An alphanumeric number is a string of alphanumeric char-
acters (excluding symbols) which starts with a number. For example,

OABC 23CAT 9XYZ

Each letter in an alphanumeric number is taken as a number, with each
letter A through Z having the value of 1 through 26 respectively, except
for E which has special meaning and is explained below.

A = 1

J = 10

S = 19

B = 2

K ■= 11

T = 20

C = 3

L = 12

U = 21

D = 4

M = 13

V = 22

E = (exponentiation)

N = 14

W- 23

F = 6

O = 15

X = 24

G = 7

P = 16

Y = 25

H = 8

Q = 17

Z = 26

I = 9

R = 18

An easy way to give FOCAL numerical valued letters is to start
with the number 0, as in the following example.

*TYPE%2,0AB

= 12*

9-12

Since after 0, A= 1 and B=2, therefore, AB= 12. Also,

♦TYPE OAB+OC

= 15*

Since after O, A=l, B=2, and C=3, then 12+3 = 15. Therefore,

♦TYPE0XYZ+1

= 2677*

because

X = 24
Y = 25
Z = 26
+ 1+1

2677
The above example can be solved using the following algorithm.

(X times 10 2 ) + (Y times 10 1 ) + (Z times 10°) + 1 = 2677
or

(24 X 100) + (25 X 10) + (26 X 1) + 1 = 2677

Taken as a numeral, the letter E has special meaning. It denotes
exponentiation, where the subsequent alphanumerics aire taken as the
exponent of the preceding alphanumerics.

♦TYPE %8, OAEC

= 1000* (AflO^ltlO 3 )

♦TYPE OAEG

= 10000000* (Afl0 c =l X 10 T )

Only one E is allowed in any one alphanumeric number.

Alphabetic characters may be used when assigning numerical values
to identifiers or variables in response to an ASK statement. An example
of this use can be found in lines 3.20 and 3.30 of "Intercept and Plot
of Two Functions."

FOCAL COMMANDS
TYPE

The TYPE command is used to request that FOCAL compute and
type out a text string, the result of an expression, or the value of an
identifier. For example

4.14 TYPE 8.1+3.2 - (29.3*5)/2.5f7

4.15 TYPE (2.2+3.5)* (7.2/ 3 )/59.1t3

9-13

Several expressions may be computed in a single TYPE command,
with commas separating each expression,

*ERASE

*9.19 TYPE %4.01, Al*2, E+2f5, 2.51*81.1

*DO 9.19

= 0.0= 32.0= 204*

The output format may be included in the TYPE statement as shown
in the example above and as explained under "Handling Text Output."
The user may request a typeout of all identifiers which he has de-
fined by typing TYPE $ and a carriage return. This causes FOCAL
to type out the identifiers with their values, in the order in which they
were defined. The $ may follow other statements in a TYPE command,
but must be the last operation on the line.

87; SET Y=55; SET C9=91

♦ERASE

SETL=33;

SET B=8"

*TYPE $

L@(00) =

33.0

B@(00) =

87.0

Y@(00) =

55.0

C 9 (00)=

91.0

Any text string enclosed in quotation marks may be included in a
TYPE command. A carriage return may replace the terminating quo-
tation mark, as shown below:

* 1.2 TYPE "X SQUARED =

A text string or any FOCAL command or group of commands may
not exceed the capacity of a Teletype line, which is 72 characters on
the ASR33 Teletype. A line may not be continued on the following
line. To print out a longer text, each line must start with a TYPE
command.

NOTE: For extremely large numbers, there may be some input/output con-
version error. For example,

*TYPE 10f616

= 0.999959E+616*
Exponent overflow is not detected:

TYPE 10|617

= 0:957418E-616

9-14

Several operations are useful in formulating output.

1. FOCAL does not automatically perform a carriage return after
executing a TYPE command. The user may insert a carriage
return/line feed, by typing an exclamation mark (!).

2. To insert a carriage return without a line feed, the user types a
number sign (#).

3. Spaces may be inserted by enclosing them in quotation marks.

4. An expression may be enclosed in question marks to avoid re-
peating it in quotes (see "Using the Trace Feature").

ASK

The ASK command is normally used in indirect commands to allow
the user to input data at specific points during the execution of his
program. The ASK command is written in the form,

*1 1.99 ASK X,Y,Z,

When step 11.99 is encountered by FOCAL, it types a colon (:). The
user then types a value in any format for the identifier, followed by a
terminator, which may be space, comma, carriage return, or ALT
MODE. FOCAL then types another colon and the user types a value
for the next identifier. This continues until all the identifiers or vari-
ables in the ASK statement have been given values,

"11.99ASKX, Y.Z
*DO 11.99

:5, :4, :3,*

where the user typed 5, 4, and 3 as the values, respectively, for X,
Y, and Z.

The ALT MODE, when used as a terminator, is nonspacing and
leaves the previously defined variable unchanged, as shown below.

*SET A-5

*ASKA

: 1 23 * (user depressed the ALT MODE Key after typing 1 23 )

♦TYPE A

= 5*

ALT MODE is frequently used when the user does not wish to change
the value of one or more identifiers in an ASK command.

9-15

*11.99ASKX,Y,Z
*D0 11.99
:5, :4, :3,*

*DO 1 1 .99 (User did not wish to enter new value for Y, so he typed
:8, ::10,* ALT MODE in response to second colon.)
♦TYPEX, Y,Z
= 8= 4= 10*

FOCAL recognizes the value when its terminator is typed. There-
fore, a value can be changed but only before typing its terminator.
This is done by typing a left arrow ( *- ) immediately after the value,
and then typing the correct value followed by its terminator. This is
the exception to the use of the left arrow, as explained under "Cor-
rections."

Text strings and format control characters may be included in an
ASK statement by enclosing the string in quotation marks.

*1.10 ASK "HOW MANY APPLES DO YOU HAVE?" APPLES

*DO 1.10

HOW MANY APPLES DO YOU HAVE? :25

The identifier AP (FOCAL recognizes the first two characters only)
now has the value 25.

Alphabetic Responses

Alphabetic characters may be used to assign numerical values to
identifiers or variables :

♦1.1 ASK A; TYPE %4, A

*DO 1.1

:ABCD = 1234*
Where the user typed ABCD and FOCAL typed the numerical value
of ABCD, as was explained under "Alphanumeric Numbers."

WRITE

A WRITE command without an argument (or with the argument
ALL) causes FOCAL to type out all indirect statements which the
user has typed. Indirect statements are those preceded by a line number.

9-16

When the user types

*WRITE
♦WRITE ALL

or simply

and the RETURN key, FOCAL types out a copy of all previously
typed indirect statements.

A specific line or group of lines may be typed out with the WRITE
command using arguments:

♦WRITE 2. 1 (FOCAL types line 2.1)

♦WRITE 2.0 (FOCAL types all group 2 lines)

SET

The SET command is used to define identifiers. When FOCAL
executes a SET command, the identifier and its value is stored in the
user's symbol table, and that value will be substituted for the identifier
when the identifier is encountered in the program.

♦ERASE ALL

♦3.4 SET A=2.55; SET B=8.05

*3.5TYPE%,A+B

♦GO

= 0.106000E+02*

An identifier may be set equal to previously defined identifiers, which
appear in arithmetic expressions.

*3.7 SET G=(A+B)*2.2f5
*

ERASE

An ERASE command without an argument is used to delete all
identifiers, with their values, from the symbol table.

If the ERASE command is followed by a group number or a specific
line number, a group of lines or a specific line is deleted from the pro-
gram.

♦ERASE 2 (deletes all group 2 lines)

♦ERASE 7.11 (deletes line 7.11)

The ERASE ALL command erases all of the user's input (i.e.,
symbol table entries and commands),

9-17

In the following example, an ERASE command is used to delete
line 1.50.

♦ERASE ALL

* 1.20 SET B=2

* 1.30 SET C=4
H.40TYPEB+C
"1.50TYPEB-C
*ERASE 1.50
♦WRITE ALL
C-FOCAL, 8/68
01.20 SET B=2
01.30 SET C=4
01.40 TYPE B+C

The GO command requests that FOCAL execute the program, start-
ing with the lowest numbered line. The remainder of the program
will be executed in line number sequence. Line numbers must be in
the range 1.01 to 15.99. The GO command cannot be given indirectly.

GOTO

The GOTO command causes FOCAL to transfer control to a specific
line in the indirect program. It must be followed by a specific line
number. After executing the command at the specified line, FOCAL
continues to the next larger line number, executing the program se-
quentially. GOTO is a single word.

ERASE ALL
*1.1TYPE"A"

* 1.2 TYPE "B"

* 1.3 TYPE "C"

* 1.4 TYPE "D"
*GOTO 1.2
BCD*

The DO command transfers control momentarily to a single line, a
group of lines, or the entire indirect program. If transfer is made to a
single line, the statements on that line are executed, and control is
transferred back to the statement following the DO command. Thus,

9-18

the DO command makes a subroutine of the commands transferred to,
as shown in this example,

♦ERASE ALL
*1.1 TYPE"X"
*1.2D0 2.3;TYPE"Y"

* 1.3 TYPE "Z"
*2.3 TYPE "A"
*GO
XAYZA*

If a DO command transfers control to a group of lines, FOCAL
executes the group sequentially and returns control to the statement
following the DO command.

If DO is written without an argument or the user writes DO ALL,
FOCAL executes the entire indirect program.

With arguments, DO commands cause specified portions of the
indirect program to be executed as closed subroutines. These sub-
routines may also be terminated by a RETURN command.

If a GOTO or an IF command is executed within a DO subroutine,
two actions are possible:

1. If a GOTO or IF command transfers to a line inside the DO
group, the remaining commands in that group will be executed
as in any subroutine before returning to the command following
the DO.

2. If transfer is to a line outside the DO group, that line is executed
and control is returned to the command following the DO; unless
that line contains another GOTO or IF.

♦ERASE ALL

* 1.1' TYPE "A"; SETX=-1; DO 3.1; TYPE "D"; DO 2

♦1.2D0 2.2

*2.1TYPE"G"
*2.2 IF (X)2.5,2.6,2;7
*2.5 TYPE "H"
♦2.6 TYPE "I"
*2.7 TYPE "J"
*2,8TYPE"K"

*2.9 TYPE %2.01, X; TYPE " "; SET X=X+1

*3.1 TYPE "B"; GOTO 5.1; TYPE "F"

9-19

*5.1TYPE"C"
*5.2 TYPE "E"
*5.3 TYPE "L"
*GO

(FOCAL types the answer)

ABCDGHIJK=— 1.0 IGIJK=0.0 BCEL*

In order to transfer control after a comparison, FOCAL contains a
conditional IF statement. The normal form of the IF statement consists
of the word IF, a space, a parenthesized expression or variable, and
three line numbers in order, separated by commas. The expression is
evaluated, and the program transfers control to the first line number if
the expression is less than zero, to the second line number if the expres-
sion has a value of zero, or to the third line number if the value of the
expression is greater than zero.

The program below transfers control to line number 2.10, 2.30, or
2.50, according to the value of the expression in the IF statement.

*2.1 TYPE "LESS THAN ZERO"; QUIT
*2.3 TYPE "EQUAL TO ZERO"; QUIT
♦2.5 TYPE "GREATER THAN ZERO"; QUIT
*IF (25-25)2.1,2.3,2.5
EQUAL TO ZERO*

The IF statement may be shortened by terminating it with a semi-
colon or carriage return after the first or second line number. If a semi-
colon follows the first line number, the expression is tested and control
is transferred to that line if the expression is less than zero. If the
expression is not less than zero, the program continues with the next
statement,

*2.20IF(X)1.8;TYPE"Q"

In the above example, when line 2.20 is executed, if X is less than
zero, control is transferred to line 1.8. If not, Q is typed out.

*3.19IF (B) 1.8,1.9
*3.20TYPEB

In this example, if B is less than zero, control goes to line 1.8; if B
is equal to zero, control goes to line 1.9; if B is greater than zero,

9-20

control goes to the next statement, which in this case is line 3.20, and
the value of B is typed. The expression must be enclosed in paren-
theses, but other enclosures may be used within the expression.

RETURN

The RETURN command is used to exit from a DO subroutine.
When a RETURN command is encountered during execution of a DO
subroutine, the program exits from its subroutine status and returns to
the command following the DO command that initiated the subroutine
status.

QUIT

A QUIT command causes the program to halt and return control
to the user. FOCAL types an asterisk and the user may type another
command. This command is suggested as the formal end of any pro-
gram.

Comment

Beginning a command string with the letter C will cause the re-
mainder of that line to be ignored so that comments may be inserted
into the program. Such lines will be skipped over when the program is
executed, but will be typed out by a WRITE command. A program
that is well documented with comments is much more meaningful and
easier to understand than one without comments.

FOR

This command is used in setting up program loops and iterations.
The general format is

FOR A=B, C, D; (command)

The identifier A is initialized to the value B, then the command follow-
ing the semicolon is executed, at least once. After the command has
been executed, the value of A is incremented by C and compared to
the value of D. If A is less than or equal to D, the command after the
semicolon is executed again. This process is repeated until A is greater
than D, and FOCAL goes to the next sequential line.

The identifier A must be a single variable. B, C, and D may be
either expressions, variables, or numbers. If comma and the value C
are omitted, it is assumed that the increment is 1. If C,D is omitted, it
is handled like a SET statement and no iteration is performed.

The computations involved in the FOR statement are done in float-
ing-point arithmetic, and it may be necessary, in some circumstances,
to account for this type of arithmetic computation.

9-21

Example 1 below is a simple example of how FOCAL executes a
FOR command. Example 2 shows the FOR command combined with
a DO command.

Example 1:

*ERASE ALL

*1.1 SETA=100

*1.2 FOR B=l,1,5; TYPE %5.02, "B IS " B+A.!

*GO

B IS = 101.00

B IS = 102.00

B IS = 103.00

B TS = 104.00

B IS = 105.00

Example 2:

*l.lFORX=l,1.5;DO2.0

*1.2 GOTO 3.1

*2.1TYPE!" "%3,"X"X

*2.2 SET A=X+ 100.000

*2.3 TYPE !" " %5.02, "A "A

'3.1

QUIT

X =

A =

101.00

X =

A =

102.00

X =

A =

103.00

X =

A =

104.00

X =

A =

105.00*

MODIFY

Frequently, only a few characters in a particular line require changes.
To facilitate this job, and to eliminate the need to replace the entire
line, the FOCAL programmer may use the MODIFY command. Thus,
in order to modify the characters in line 5.41, the user types MODIFY
5.41. This command is terminated by a carriage return whereupon the
program waits for the user. to type that character iq the position in
which he wishes to make changes or additions. This character is not

9-22

printed. After he has typed the search character, the program types out
the contents of that line until the search character is typed.

At this point, the user has seven options:

1. Type in new characters in addition to the ones that have already
been typed out.

2. Type a form-feed (CTRL/L) ; this will cause the search to pro-
ceed to the next occurrence, if any, of the search character.

3. Type a CTRL/BELL; this allows the user to change the search
character just as he did when first beginning to use the MODIFY
command.

4. Use the RUBOUT key to delete one character to the left each
time RUBOUT is depressed.

5. Type a left arrow ( «- ) to delete the line over to the left margin.

6. Type a carriage return to terminate the line at that point, re-
moving the text to the right.

7. Type a LINE FEED to save the remainder of the line.

The ERASE ALL and MODIFY commands are generally used only
in immediate mode since they return to command mode upon com-
pletion. (The reason for this is that internal pointers may be changed
by these commands.)

During command input, the left arrow will delete the line numbers
as well as the text if the left arrow is the right most character on the
line. During MODIFY the line numbers cannot be changed.

Notice the errors in line 7.01 below.

•7.01 JACK AND BILL W$NT UP THE HALL

♦MODIFY 7.01

JACK AND BVILL W$\ENT UP THE HAMLL

*WRITE 7.01

07.01 JACK AND JILL WENT UP THE HILL
*

To modify line 7.01, a B was typed by the user to indicate the char-
acter to be changed. FOCAL stopped typing when it encountered the
search character, B. The user typed the RUBOUT key to delete the B,
and then typed the correct letter J. He then typed the CTRL/BELL
keys followed by the $, the next character to be changed. The RUB-
OUT deleted the $ character, and the user typed an E. Again a search
was made for an A character. This was changed to I. A LINE FEED
was typed to save the remainder of the line.

9-23

NOTE

When the MODIFY command is used the values in the user's symbol
table are reset to zero. Therefore, if the user defines his symbols in direct
statements and Ihen uses a MODIFY command, the values of his symbols
are erased and must be redefined.

However, if the user defines his symbols by means of indirect statements
prior to using a MODIFY command, the values will not be erased because
these symbols are not entered in the symbol table until the statements de-
fining them are executed.

Using the Trace Feature

The trace feature is useful in checking an operating program. Those
parts of the program which the user has enclosed in question marks
will be printed out as they are executed by FOCAL.

In the following example, parts of three lines will be printed.
♦ERASE ALL
•1.1 SETA=1
*1.2SETB=5
*1.3SETC=3
*1.4 TYPE %2, 7A+B-C?,!
* 1.5 TYPE 7B+A/C?,!
*1.6TYPE?B-C/A?
*GO

A+B-C= 3
B+A/C= 5
B-C/A= 2*
When only one ? is inserted the trace feature becomes operative
. when FOCAL encounters the ? during execution, and the program is
printed out from that point until another ? is encountered (the pro-
gram may loop through the same ?), until an error is encountered
(execution stops and an error message is typed), or until program
completion.

•ERASE ALL

*1.1 ?SET A=OB; TYPE %3, A!

* 1.2 FOR B= 1,1, 4; TYPE B+ A!

*GO

SET A=0B; TYPE % 3, A!

= FOR B=l, 1,4; TYPE B+A!

= 1 TYPE B+A!

= 2 TYPE B+A!

= 3 TYPE B+A!

= 4*

9-24

In this example, FOCAL encountered the ? as it entered line 1.1
and traced the entire program. In the following example an error has
been inserted in the FOR statement— FOCAL will detect it.

♦ERASE ALL

* 1 . 1 SET A=0B; TYPE % 3, A!

•1.2 FOR B=l,l :4; TYPE B+A!

*GO?

C-FOCAL SET A=0B; TYPE % 3, A!

FOR B= 1,1 :?07. 14 ® 01.20

GO? traced the entire program under the same conditions as explained
above when only one ? is inserted.
Mathematical Functions

Functions are provided to give extended arithmetic capabilities and
to give the potential for expansion to additional input/output devices.
A standard function call consists of four letters beginning with the letter
F and followed by a parenthetical expression.
FSGN(A-B*2)
There are three basic types of functions, two of which are included
in the basic FOCAL program.

The first type contains integer part, sign part, and absolute value
functions.

The second type, the extended arithmetic functions, are loaded at
the option of the user. They will consume approximately 800 locations
of the users program storage area. These arithmetic functions are
adapted from the extended arithmetic functions of the PDP-8 three-
word floating-point package and are fully described in the Floating
Point System manual, Order No. Digital-8-5-S.

The third type are the input/output functions; These include a non-
statistical random number generator (FRAN). This function uses the
FOCAL program itself as a table of random numbers. An expanded
version could incorporate the random number generator from the
DECUS library. Following are examples of the functions now available.
1. The square root function (FSQT) computes the square root of
the expression within parentheses.
*TYPE %2, FSQT(4)
= 2*

*TYPEFSQT(9)
= 3*

*TYPE FSQT(144)
= 12* '

9.25

2. The absolute value function (FABS) outputs the absolute or
positive value of the number in parentheses.

♦TYPE FABS(-66)

= 66*

TYPE FABS (-23)

= 23*

*TYPE FABS (-99)

= 99*

3. The sign part function (FSGN) outputs the sign part (-f- or — )
of a number and the integer part becomes a 1.

*TYPE FSGN(4-6)

=- 1*

*TYPE FSGN (4-4)
_ !*

TYPE FSGN (-7)

=- 1*

4. The integer function (FITR) has the value of the integer part
of any number:

*TYPE FITR(5.2)

= 5*

It may be used for truncating:
*TYPE FITR(55.66+.5)
= 56*

♦TYPE FITR(77.434+.5)
= 77*

For negative numbers, FITR gives the next smaller integer:
*TYPE FITR(~4.1)
=- 5*

5. The random number generator function (FRAN) computes a
nonstatistical pseudo-random number between ±1.

*TYPE%,FRAN( )
=— 0.250000E+00*
*TYPEFRAN( )
=— 0.623535E+00*

9-26

6. The exponential function (FEXP) computes e to the power
within parentheses. (e=2.718281)

♦TYPE FEXP(6.66953E— 1)
= 0.194829E+01*
TYPE FEXP(.666953)
= 0.194829E+01*
*TYPEFEXP(1.23456)
= 0.343687E+01*
TYPE FEXP (-1.)
= 0.367879E+00*

7. The sine function (FSIN) calculates the sine of an angle in
radians,

•TYPE %,FSIN(3. 14159)
= 0.238419E— 05*
•TYPE FSIN(1. 400)
= 0.985450E+00*

Since FOCAL requires that angles must be expressed in radians, to find
a function of an angle in degrees, the conversion factor, tt/180, must
be used. To find the sine of 15 degrees,

*SETP1=3.14159; TYPE FSIN(15*PI/180) .

= 0.258819E+00*

•TYPE FSIN(45*3.14159/ 180)

= 0.707106E+00*

8. The cosine function (FCOS) calculates the cosine of an angle
in radians.

•TYPE FCOS(2*3. 141592)

= 0.100000E+01*

•TYPE FCOS(.50000)

= 0.877582E+00*

•TYPE FCOS(45*3.141592/ 180)

= 0.707107E+00*

9-27

9. The arc tangent function (FATN) calculates the angle in radians
whose tangent is the argument within parentheses.

*TYPEFATN(1.)
= 0.785398E+00*
♦TYPE FATN(.3 1305)
= 0.303386E+00*
♦TYPE FATN(3. 141592)
= 0.126263E+01*

10. The logarithm function (FLOG) computes the natural loga-
rithm (logs) of the number within parentheses.

*TYPE FLOG(l.OOOOO)

= 0.000000E+00*

*TYPE FLOG (1.98765)

= 0.686953E+00*

*TYPE %5.03, FLOG(2.065)

= 0.725*
Reading From the High-Speed Reader

Up to this point we have considered only one source of input to
FOCAL programs — the Teletype keyboard. We learned that when
FOCAL types an *, it is ready to accept user commands.

The user can switch the input device to the high-speed paper tape
reader by typing an asterisk in the first position, or immediately follow-
ing the line number in a statement. The following statements cause
FOCAL to read a tape from the high-speed reader.

* * ; The first asterisk is that typed out by FOCAL, inform-

ing the user that a command may be input from
the keyboard. The second asterisk typed by the
user, switches input to the high-speed reader.

*1.21* The user types * after the line number.

To switch back to the keyboard, the user types another *. If there
is no tape in the high-speed reader, or when an "end of tape" condition
is reached, FOCAL automatically switches back to keyboard input.

9-28

This feature is useful for loading FOCAL programs and for inputting
large amounts of data during execution of a FOCAL program.

When the following statement is executed FOCAL accepts four
pieces of data from the high-speed reader.

*1.10*; FOR 1=1,4; ASK HR(I)

*i:n*

♦DO 1.10

The user typed an asterisk after line 1.11 to return control to the key-
board. If the tape contains fewer than four pieces of data, the remain-
ing pieces will be taken from the keyboard.

Generating Program Tapes

To generate a program tape of the user's FOCAL program on the
low-speed paper tape punch, the user should

1. Respond to * by typing WRITE ALL (do not depress the RE-
TURN key).

2. Depress low-speed punch to ON.

3. Generate leader tape (depress the SHIFT, REPT, and P keys
in that order; release in reverse order) .

4. Depress RETURN key.

When the user's FOCAL program has been typed and punched, the
user should

5. Generate trailer tape (depress the SHIFT, REPT, and P keys
in that order; release in reverse order) .

6. Depress low-speed punch to OFF.

7 . Remove and label the punched paper tape.

The user's FOCAL program is still in the computer. FOCAL is
waiting for user input.

9-29

EXAMPLES OF FOCAL PROGRAMS

The following programs reveal some of FOCAL's features in various
applications. The examples show that FOCAL finds practical applica-
tion in any situation:

FORTRAN-type problems are handled easily with little programming

time.

FOCAL's easy-to-leara language allows the user to concentrate more

on his problem than on programming.

FOCAL as a tool is easier to learn and use than a slide rule or any

other desk calculator, and it offers vastly more than any combination

of previous problem solving tools.

FOCAL can calculate complex problems and print and/ or display

the results in "one fell swoop."

The example programs included in this section are maintained on
punched paper tape. Each program was loaded into core using
FOCAL's high-speed paper tape reader input feature. The WRITE
command was then used to get the program printout for inclusion here.
Then the GO command was issued to execute each program.

When using the WRITE command, FOCAL immediately identifies
the version of the FOCAL tape being used — in this case, C-FOCAL,
8/68. The C preceding FOCAL , 8/68 is the comment line indicator.

Table Generation Using Functions

The ability to evaluate simple arithmetic expressions and to generate
values with the aid of library functions is one of the first benefits to be
obtained from learning the FOCAL language. In this example, a table
of the sine, cosine, natural logarithm, and exponential values is gen-
erated for a series of arguments. As one becomes familiar with these
and other library functions, it becomes easy to combine them with the
standard arithmetic operations of addition, subtraction, multiplication,
division, and exponentiation. The user is then able to evaluate any given
formula for a single value or for a range of values as in this example.

Although FOCAL allows the typing of more than one command per
line, each command in this example has been typed on a separate line
to maintain clarity and because of the length of several of the com-
mands. In this example, line 01.05 outputs the desired column head-
ings. Line 01.10 is the loop to generate values for I, beginning with the
value 1.00000 and continuing in increments of .00001 up through the
value 1.00010; the DO 2.05 command at the end of this second line
causes line 02.05 to be executed for each value of I. Line 02.05 is the
command to evaluate the various library functions for the I arguments;
the %7.06 specifies that all output results up to the next % symbol

9-30

are to. appear iiTfixed-point format with one digit position to the left of
the decimal and six digit positions to the fight: the second % symbol
reverts the output mode back to floating point for the remaining values
— FLOG(I). and FEXP(I). Line 01.20 (optional) returns control
to the user.

Several techniques can be noted in line 02.05 of this example.

1. FOCAL commands can be abbreviated to the first letter of the
command followed by a space, as shown by the use of T instead
of TYPE. This technique can be used to shorten command
strings.

2. Arguments can be enclosed in various ways: ( ), < >»[.]■
This ability is useful in matching correctly when a number of
such enclosures appear in a command.

3. Spaces can be inserted in an output format by enclosing the ap-
propriate number of spaces within quotation marks. Such use of
spacing is recommended to improve the readability of the output
results.

4. FOCAL presently allows accuracy of six significant digits, which
makes possible the use of very small loop increments (in this
example, .00001); this should eliminate the need to interpolate
between table values of trigonometry functions in most cases.

C-l

FOCAL , 8/68

.0 5 T "

SINE

COSINE

LOG E"!

.10 FOR I

= 1 » .00001, 1 .

000U DO 2.05

.20 QUIT

• 05 T 7.7.

06, I»" ",FS1N(D," ",FC0S«I>," ", %, FLCIGC I ]> " ", FEXP< hi , 1

*G0

SINE

COSINE

LOG

1 .000000

= 0.841471

0.540 303

.000000E+00

0.271828E+01

1 .000010

= 0.841476

0.540295

0. 97750 7E-0S

0.271831E+01

1 .000020

= 0,841481

0.540286

0.195501E-04

0.271834E+01

1 .000030

= 0.e4l487

0.540278

0.293249E-04

0.271836E+01

1 .000040

= 0.841492

0.540270

0.390997E-04

0.271839E+01

1 .000050

= 0.841497

0.540262

0.488744E-04

0.271842E+01

1 .000060

= 0.841503

0.540254

0.586490E-04

0.271844E+01

1 .0000 70

= 0.841508

' =

0.540245

0.684235E-04

0.271847E+01

1 .000080

= 0.841513

0.540237

0.781980E-0 4

' =

0.271849E+01

1 .000090

= 0.841519

0.540229

0.879 723E-04

0.271852E+01

1 .000100

= 0.841524

0.540221

0.977465E-04

0.271855E+01

Addition Exerciser Using Functions

FOCAL randomly selects pairs of 1- or 2-digit positive integers and
sets them up in an addition problem. The user types the answer to the
addition problem. FOCAL then checks the user's answer against its
own and tells the user whether he was right or wrong. If the user's
answer is correct, FOCAL types another addition problem. If the user's

9-31

answer is wrong, FOCAL gives him two more tries at the problem. If
after three tries the user still has not typed the correct answer, FOCAL
suggests that the user consult his teacher, gives the correct answer to
the problem, and goes on to another problem.

This program can be used as a drill for the young student of addition,
and with a few modifications it can be extended to give only subtraction
problems or randomly vary between addition and subtraction.

The program uses three functions in line 01.10.

C-FOCAL , 8/68

01.05 TYPE "HELLO, PLEASE ADD THE FOLLOWING SETS OF NUKBERS."!
01.10 SET A=FABSCFITR«100*FRANU»>J SET B=FABS<FITR<99*FRANC ]»>

01.20 TYPE %7, A. IB, 1 " "I

01.30 ASK REPLY.!

01.40 IF CREPLY-A-B) 2.1,1.5,2.1

01.50 SET WR=0JTYPE "THAT IS CORRECT."!

01.60 GOTO 1 »1

02.10 SET WR = WR+U IF CWR-2) 2.2,2.2,3.1
02.20 T "SORRY, TRY AGAIN, "1 J GOTO 1.2

03.10 T "IF YOU ARE HAVING TROUBLE, ASK YOUR TEACHER FOR HELP."!

03.20 TYPE "THE CORRECT ANSWER IS "A+B, I

03.30 GOTO 1 .1

♦GO

HELLO, PLEASE ADD THE FOLLOWING SETS OF NUMBERS.

«1

SORRY

, TRY

AGAIN,

THAT

IS CORRECT.
32

S32

THAT

IS CORRECT.
28

S33

THAT IS CORRECT.

9-32

= 8

= A

184

SORRY. TRY

AGAIN.

" 8

■ A

SORRY. TRY

AGAIN.

■ 8

= A

:48

IF YOU ARE HAVING TROUBLE. ASK YOUR TEACHER FOR HELP.
THE CORRECT ANSWER IS = 12

7
=

!10

IF YOU ARE HAVING TROUBLE. ASK YOUR TEACHER FOR HELP.
THE CORRECT ANSWER IS = 7
= 6
= 81

187

THAT

CORRECT.

112

THAT

CORRECT.

«17

SORRY,

TRY AGAIN,
9
7

116

THAT IS CORRECT.

9-33

Finding Roots of a Quadratic Equation

This program uses the square root function to analyze a quadratic
equation of the form:

Y = AX 2 +BX+C

The analysis is followed by output of the roots and/or a comment
concerning the nature of the roots. After the output the program re-
starts and accepts new values for A, B, and C. Some interesting fea-
tures are:

1. It is not phased by a negative discriminant.

2. Input is set up so that if the space bar is used as a terminator,
A, B, and C will be printed in a row format which is suggestive
of their format in the equation.

3. If A is equal to zero, FOCAL types an error message. There-
fore, division by zero is not possible.

C-FOCAL , 8/68

01.10 ASK ! 7A B C?J SET RQ0T=B«2-4*A*C

01 .20 IF CA> 1 .4,1 .3,1.4

01.30 TYPE ! "THIS IS A FIRST DEGREE EQUATION" U GOTO 1.10

01.40 TYPE X6.03, I " THE ROOTS ARE"* IF CROOT) 1.7,1.6

01.50 TYPE 1,<-B+FSQT<R00T>)/'2*A, ! , C-B-FSQT<ROOT>) /2*AJ GOTO 1.1

01.60 TYPE 1 -B/2*A,!J GOTO 1.10

01.70 TYPE " IMAGINARY"!, -B/2*A," + (", FSQTC-ROOT) /2*A,">" "»!•

01.80 TYPE !»-B/2*A," - C",FSQT[-R00T3/2*A,") *I ", !j GOTO 1.10

♦ GO

A }4 B :2 Cl8

THE ROOTS ARE IMAGINARY
=- 0.250 + (= 1.392)*I

=- 0.250 - <= 1.392)*1

A 10 B J2 CI4

THIS IS A FIRST DEGREE EQUATION

A II B S4 CJ4

THE ROOTS ARE
=- 2.000

A si B «-2 Cs4

THE ROOTS ARE IMAGINARY

= 1.000 + (= 1.732)*I

= 1.000 - (= 1.732>*I

A «4 B S6 C:8

THE ROOTS ARE IMAGINARY
=- 0.750 + <= 1.199)#I
=- 0.750 - (= 1 . 1995*1

A II B 16 C:8

THE ROOTS ARE
=- 2.000
=- 4.000

9-34

Square Completer Using the FOR Command

The program directs the FOCAL user to specify the values for A,
B, and C for the paraholic function of the form:

F(X)=AX2 + BX+C

The program quickly manipulates these values and prints an expression
of the same function in the form :

F(X) = A(X-K)2 + C-K2

where in line number 01.20, below, K is SET equal to — B/2A, the
value of the parabola's axis of symmetry.

It is brought out in many high school algebra classes that this is a
handy way to visualize a parabolic function since it expresses, in easily
readable forms, various characteristics of the graph of the function.
The program is designed to do the complicated arithmetic involved in
completing the square of a parabolic function.

C-FOCAL , 8/68

01.09 TYPE !, "INPUT A,B,C SUCH THAT FCX) =2X t2 + BX+C"» !

01.10 ASK ?A B C?J IF CA) 1.2,1.4)
01.20 SET K=-B/2*AI SET C=C/A-K»2

01.30 TYPE X4.02, !»" F<X)",A."*«X-",K»") tS",C ! J GOTO 1.09
01.40 TYPE i, "THAT IS NO QUADRATIC!",!!) GOTO 1.09
*
♦ GO .

INPUT A,B,C SUCH THAT FCX)=2X»2+BX+C
A !l B :l CM
F(X>= 1.00*(X- = - 0.50X2= 0.75
I
INPUT A,B,C SUCH THAT FCX) =2X T2+BX+C
As3Bs6Cs3
•FCX)= 3.00*(X-=- 1.00X2 = 0.00

INPUT A»B,C SUCH THAT F<X).=2X t2 + BX+C
A :3 B !18 CI24
FCX> = 3.00*CX- = - 3.00X2 = -. 1 .00

INPUT A,B,C SUCH THAT FCX) =2X '2+BX+C
A U B U Csl2
FCX>= 1.00*CX- = - 0.50X2= 11.75

INPUT A,B»C SUCH THAT F<X> =2X »2+BX+C
A :AB B !CD C:FG
FCXX 12.00*<X- = - 1.42X2= 3.58

INPUT A»B,C SUCH THAT F CX) =2X f 2+BX+C

A :0 B :l CsB

THAT IS NO QUADRATIC!

INPUT A,B,C SUCH THAT F CX) =2X t2+BX+C
A !2 B :4 C:6
FCX)= 2.00*(X- = - 1.00X2= 2.00

9-35

Interest Payment Program

This is an example of a business-oriented FOCAL program. It is
designed to give a complete picture of the payments which will be made
on a loan, with interest, on an installment plan basis.

Under program control, the computer requests as input the amount
of a loan, the percentage of interest on that loan, and the length of
time over which the loan is to be paid. The computer then calculates
and types the amount of monthly payments to be paid, the total amount
of interest which will be paid, and a table showing interest paid, amount
applied to principle, and balance due after each payment,

C-FOCAL . 8/68

01.10 ASK I 1X7.02, "ENTER INTEREST IN PERCENT

01 .14 SET J=J/100

01-16 ASK "ENTER AMOUNT OF LOAN " A* 1

01.20 ASK "NUMBER OF YEARS " N» !

01.24 ASK "NUMBER OF PAYMENTS PER YEAR " M, ! I

01-30 SET N=N*MJ SET I=J/M

01 .34 SET B=l+I

01.40 SET R=A*I/C!-1/BTN>

01.42 TYPE "MONTHLY PAYMENT " R, I

01.48 TYPE "TOTAL INTEREST " R*N-A, ! !

J»!

02.05

SET B=A

02. t0

TYPE " INTEREST APP TO PRIN

02.12

SET L=B*lJ SET P=R-L

02.16

SET B=B-P

02.18

TYPE L» " "P." "B,

02.20

IF CB-R) 2.24.2.24,2.12

02.24

TYPE B*I," "R-B»I,t "LAST PA'

02.30

*
*G0

QUIT

BALANCE", I

B*I+B, I

ENTER INTEREST IN PERCENT i5
ENTER AMOUNT OF LOAN 12345.00
NUKBER OF YEARS S4
NUMBER OF PAYMENTS PER YEAR :3

MONTHLY PAYMENT

217.23

TOTAL INTEREST

261 .75

INTEREST

APP

TO PRIN

39.08

176.15

36.12

181.12

33.10

184.13

30.03

187.20

26.91

190.32

23.74

193.50

20.51

196.72

17.23

200.00

13.90

203.33

10.51

206.72

7.06

210.17

3.56

213.67

LAST

PAYMENT! =

21 7.21

BALANCE

2166.85

1985.74

1801 .61

1614.40

1424.08

1230.58

1033.86

833.87

630.53

423.81

213.65

9-36

Temperature Conversion Using the FOR Command

The ability for loop parameters to be negative, zero, fractional, or ex-
pressions, provides power beyond many other similar languages in
simplifying the routine's structure. It also reemphasizes the flexibility
and control over FOCAL programs at the time they are run.

Measurement system conversions are time consuming in many lines
of work. A short FOCAL program, such as the one illustrated in the
following example, eliminates hours of repeated calculations. In this
particular example, the problem is to convert temperatures from de-
grees Fahrenheit to degrees Centigrade, using the formula:

T°C=5/9(T°F-32)

This routine is quite similar in structure to the "Table Generation"
example. The one basic difference is that here the user can input the
loop parameters which govern the generation of the output. Thus, pro-
vision has been made for output of properly labeled requests for start-
ing, ending, and incrementing values and their input for use by the
program.

C-FOCAL .8/68

02.10 ASK I, "FROM ".START," TO ".END," DEGREES FAHRENHEI T", I

02.20 ASK " IN INCREMENTS OF ",INCR," DEGREES",!!

02.30 TYPE "THE APPROXIMATE FAHRENHEIT TO CENTIGRADE CONVERSIONS ARE!'

02-40 FOR T=START,INCR,END3 TYPE U DO 2.50

02.45 QUIT

02.50 TYPE " ",T," FAHRENHEIT DEG. ", <T-32>*5/9," CENTIGRADE DEG-"

*G0

FROM !-40 TO 180 DEGREES FAHRENHEIT
IN INCREMENTS OF 120 DEGREES

THE APPROXIMATE FAHRENHEIT TO CENTIGRADE CONVERSIONS AREi

40.00 FAHRENHEIT DEG. =- 40.00 CENTIGRADE DEG.

20.00 FAHRENHEIT DEG. =- 28.89 CENTIGRADE DEG.

= 0.00 FAHRENHEIT DEG. »- 17.78 CENTIGRADE DEG.

= 20.00 FAHRENHEIT DEG. =- 6.67 CENTIGRADE DEG.

= 40.00 FAHRENHEIT DEG. » 4.45 CENTIGRADE DEG.

■ 60.00 FAHRENHEIT DEG. = 15.56 CENTIGRADE DEG.

* 80.00 FAHRENHEIT DEG. = 26.67 CENTIGRADE DEG.*

9-37

One-Line Function Plotting

This example demonstrates the use of FOCAL to present, in, graphic
form, some given function over a range of values. In this example, the
function used is

y = 30+15(SIN(x))e-.lx

with x ranging from to 15 in increments of .5. This damped sine
wave has many physical applications, especially in electronics and
mechanics (for example, in designing the shock absorbers of a car) .

In the actual coding of the example, the variables I and J were used
in place of x and y, respectively; any two variables could have been
used. The single line 08.01 contains a set of nested loops for I and J.
The J loop types spaces horizontally for the y coordinate of the func-
tion; the I loop prints the * symbol and the carriage return and line
feeds for the x coordinate. The function itself is used as the upper limit
of the J loop, again showing the power of FOCAL commands.

The technique illustrated by this example can be used to plot any
desired function. Although the * symbol was used here, any legal
FOCAL character is acceptable.

C-FOCAL , 8/68

08.01 F I=0».5j15j T "*", H F J=0,30+15*FSINCI)*FEXPC-.1*I]J T " "

*DO 8.01

*
*

*
#

9-38

Intercept and Plot of Two Functions

Values are first computed and printed for two monotonic functions.
Then these curves are plotted within specified limits. Nonmonotonic
functions must be plotted using the method of residuals.

C-FOCAL , 8/68

01.02 ASK "LOWER LI MI T",LL, ! "UPPER LI MI T"» UL, ! "I NCREMENT", I N, !

01 .10 SET Y1=0J SET Y2=0J

01-20 FOR X=LL,IN,ULJ SET Y1=-X-3J SET Y2=3+4*X-X t2 J DO 2.0

02.10 IF (Y2-Y1) 2.4,2.2.2.4

02.20 TYPE !! "THE POINT OF INTERSECTION IS ", ! J GOTO 2.3

02.30 TYPE "XI", X." ","Y1",Y1, ! , "X2", X, " . ", "Y2"# Y2, ! ! 3 RETURN

02.40 TYPE "XI", X," ","Y1",Y1,!,"X2",X," ", "Y2", Y2, ! !

03-10 TYPE "DO YOU WANT A PLOT?"

03.20 ASK "(TYPE Y FOR YES. TYPE N FOR NO) ",AN, !!

03.30 IF CAN-25) 9.1,4.1

04.10 FOR X=LL,IN,ULi DO 5.0

05.01 IF CX> 5.1,5.02,5.1

05.02 TYPE " Y

05.10 FOR Y=0,30J TYPE " "

05.20 TYPE ".",#

05.30 FOR Y = 0,30+C.-X-3> J TYPE " " •

05.40 TYPE "*",#

05.50 FOR Y=0,30+C3+4*X-X'2>) TYPE

05.60 TYPE "*",!

05.70 RETURN

09.10

QUIT

*GO

LOWER LIMIT:-10

UPPER LIMIT: 10

INCREMENT: 1

Xl=- 10.00 Yl=

X2=- 10.00 Y2=-

7.00
137.00

Xl=-
X2=-

9.00.
9.00

Yl =
Y2 = -

6.00
1 1 4.00

Xl=-

X2=-

8.00
8.00

Yl =

Y2 = -

5.00
93.00

Xl=-
X2=-

7.00
7.00

Yl =
Y2 = -

4.00
74.00

Xl=-
X2 = -

6.00
6.00

Yl =

Y2 = -

3.00
57.00

Xl=-

X2 = -

5.00
5.00

Yl =

Y2 = -

2.00
42.00

Xl=-
X2=-

4.00
4.00

Yl =
Y2 = -

1 .00
29.00

Xl=-
X2 = -

3.00
3.00

Yl =

Y2 = -

0.00
18.00

Xl=-
X2 = -

2.00
2.00

Yl=-
Y2 = -

1 .00
9.00

9-39

THE POINT OF INTERSECTION IS
X1=- 1.00 Yl=- 2.00
X2=- 1 .00 Y2 = - 2.00

Xl =

.00

Yl=-

3.00

X2 =

.00

Y2 =

3.00

Xl =

1.00

Yl=-

4.00

X2 =

1 .00

Y2 =

6.00

Xl =

2.00

Yl=-

5.00

X2 =

2.00

Y2 =

7.00

Xl =

3.00

Yl=-

6.00

X2 =

3.00

Y2 =

6.00

X! =

4.00

Yl=-

7.00

X2 =

4.00

Y2 =

3.00

Xl =

5.00

YI=-

8.00

X2 =

5.00

Y2 = -

2.00

THE POINT OF INTERSECTION IS

Xl =

6.00

Yl=-

9.00

X2 =

6.00

Y2 = -

9.00

xt =

7.00

Yl=-

10.00

X2 =

7.00

Y2=-

18.00

Xl =

8.00

Yl=-

1 1 .00

X2 =

8.00

Y2 = -

29.00

Xl =

9.00

Yl=-

12.00

X2=

9.00

Y2=-

42.00

Xl =

10.00

Yl=-

13.00

X2 =

10.00

Y2 = -

57.00

Xl =

1 1 .00

Yl=-

13.00

X2 =

. 1 1 .00

Y2 = -

57.00

DO YOU WANT A PLQT7CTYPE Y FOR YES. TYPE N FOR NO) :Y

9-40

Plotting on the Oscilloscope

This is an example of using the FDXS and FDIS functions for plot-
ting on the oscilloscope. The functions are used in the SET command
of statement number 01.40 as shown below, which is equivalent to

SET H=FDXS(X)-FDIS(Y)

where X and Y are the x and y coordinates of the point to be plotted
on the scope.

The program will plot a sine wave with a user-determined number
of complete cycles (Q) .

C-FOCAL , 8/68

01.10 ASK "NUMBER OF CYCLES" Q

01.20 F 1=0. ,50* S X<I>=20*I* S YU> = <FSINC3.14159*I/<25/Q>J+l>*500

01.30 TYPE ! "READY TO PLOT" , !

01.40 FOR 1=0. ,50* SET H = FDXS<X[I3>-FDTS<Y<I»>

01 .50 GOTO 1.4

♦ GO

NUMBER OF CYCLES: 2

READY TO PLOT

Formula Evaluation for Circles and Spheres

In this example, FOCAL is used to calculate, label, and output the
following values for an indefinite number of radii typed in by the user.

Given: radius(R)

Program calculates: circle diameter 2R

circle area tR 2

circle circumference 2wR

sphere volume 4/3wR 3

sphere surface area 4*rR 2

Although the American system of inches is used in this example,
conversions to other systems (metric, for example) could be very easily
incorporated into the program, thus eliminating any need for hand-
calculated conversions.

The program is very straightforward. ASK is used to allow the user
to type in the radius value to be used in the calculations. SET is used
to supply the value of ir (PI). TYPE is used for all calculations and
output. Note that if a value (such as PI in this example) is to be
entered once and then used in repeated calculations, it should be
entered by a SET command which is outside the calculation loop,
otherwise, the variable would be set at the beginning of each pass
through the loop. However, if the value of the variable changes during
each iteration, then it must be calculated either by a SET or TYPE
command within the loop.

9-41

The use of the GOTO command (line 01.60) results in an infinite
loop of lines 01.10 through 01.60. This technique is used when the
number of desired repetitions is not known. The looping process can
be terminated at any time by typing CTRL/C. If, however, the num-
ber of desired repetitions is known (e.g., 10), the following method
can be used.

*SETPI=3.14159
•1.1 ASK ...

*1.6 TYPE ! ! ! ! !
*FOR 1=1,10; DO 1

(Eliminate GOTO 1.1)

(Direct command; causes all
steps in group 1 to be ex-
ecuted 10 times)

The ability to choose between these methods provides great flexibility
in actually running FOCAL programs.

C-F0CAL , 8/68 .

01 .01 SET PI=3. 141592
01.10 ASK "A RADIUS OF", R,

INCHES"

01 .20 TYPE

01.21 TYPE
01 .30 TYPE
01 .35 TYPE
01.40 TYPE

01.49 TYPE

01.50 TYPE

4, !, " GENERATES A CIRCLE OF:", !
DIAMETER", 2*R, " INCHES", !
AREA", PI*R»2, " SQUARE INCHES", !
CIRCUMFERENCE", 2*PI*R, " INCHES", !
AND A SPHERE OF:", 1

VOLUME", C4/3)*PI*Rt3, " CUBIC INCHES", !
AND SURFACE AREA", 4*PI*Rt2, "SQUARE INCHES"

01.60 TYPE !!!» GOTO 01.10

*GO

A RADIUS 0F:l .414 INCHES
GENERATES A CIRCLE OF:

DIAMETER= 2.8280 INCHES
AREA= 6.2813 SQUARE INCHES
CIRCUMFERENCE= 8.8844 INCHES

AND A SPHERE OF:

VOLUME= 11.8423 CUBIC INCHES

AND SURFACE AREA= 25.1252 SQUARE INCHES

A RADIUS OF:

9-42

Simultaneous Equations and Matrices

Many disciplines use subscripted variables for vectors in one, two,
or more dimensions to store and manipulate data, A common use is
the 2-dimensional array or matrix for handling sets of simultaneous
equations. For example,

Given: lX t + 2X 2 + 3X 3 = 4

4X 1 4-3X 2 + 2X S = 1

lXj + 4X 2 + 3X 3 = 2

Find: the values of X,, X 2 , and X 3 to satisfy all three equations simul-
taneously.

The solution can be reduced to simple mathematics between the
various elements of the rows and columns until correct values of X are
found, as shown in Example 3 below.

Since FOCAL uses only a single subscript, the handling of two or
more dimensions requires the generation of ajinear subscript which
represents the correct position if it were stored in normal order; i.e.,
leftmost subscript moving fastest.

Element D could be represented as
ARRAY(3); any element in this array
can be represented by a subscript in the
range through 4. The first element in
an array always has a subscript of 0.

NE DIMENSION

ARRAY( )

IN TWO DIMENSIONS

ARRAY(row, column) or A(I,J)

This must be reduced to the form A(G), where G is a function of I
and J; that is, A(I,J) = A(G). Consider the diagram

J =

1*2
1 =
1
2
3
4

The numbers along the outside edges of the box above are the 2-dimen-
sional subscripts; the numbers inside the box are the linear subscripts.

9-43

Thus each combination of I and J can be given a unique value, e.g., for
1=2 and J= 1 the element is 7.

Notice that for a constant I, increasing the value of J by one in-
creases the value of the linear subscript by five. Similarly, for a constant
J, increasing the value of I by one increases the linear subscript by
three.

The array, above, has five rows and three columns, so two values
can be defined:

IMAX = 5 and JMAX = 3

The total number of elements is IMAX * JMAX = 15. To generate
the number G in any box, using the corresponding values of I and J, the
formula is

G =1 + IMAX * J or A(G)

which is equivalent to A(I+IMAX*J). The example of solving simul-
taneous equations, above, uses this algorithm for subscripts merely by
replacing I, IMAX, and J with J, L, and K, respectively, so as to form
the equation

A(J+L*K)
Each element in a 2-dimensional array represents an area.
IN THREE DIMENSIONS

ARRAY(row, column, plane) = A(I,J,K) = A(G)

In a 3-dimensional array, each array represents a volume. Three
dimensions can be illustrated as a cube.

9-44

This cube has dimensions of five rows, three columns, and five planes;
thus, IMAX = 5, JMAX = 3, and KMAX .= 5. Each plane is num-
bered exactly as in the 2-dimensional example, except with the addition
of 15 times K (with K = the number of planes back from the first) to
each subscript in the first plane.
For example,

Upper lefthand square, back one plane from the first=15
1=0, J= 0, K=l; I+(IMAX*J) + (IMAX*JMAX*K)=15=G

or
A(0,0,1)=A(15)

IN FOUR DIMENSIONS

ARRAY (row, column, plane, cube)=A(I,J,K,L)=A(G)

Assign the values for IMAX, JMAX, KMAX; a method similar to the
one used above yields

G=I+ (IMAX*J) + (IMAX* JMAX*K) +(IMAX*JMAX*KMAX*L)

This process can be extended indefinitely to n-dimensions!

9-45

Example 1;

♦FOR J=0,4J TYPE X2.U FOR I=0,2J TYPE J+S*I

=, 0= 5= 10

= 1= 6= 11

= 2= 7= 12

■ 3= 8= 13

= 4 = 9 = 14*

Example 2:

C-FOCAL » 8/68

01.05 TYPE "ENTER 3 ROWS AND 4 COLUMNS OF NUMBERS."!

01.10 FOR J=0,2J TYPE !l FOR K=0,3J ASK N0CJ+3*K)

01.15 SET MAX=NOC0>

01.20 FOR J=0,2J FOR K=0.»3J DO 02.00

01.25 TYPE I, "LARGEST NUMBER IS " , MAXJ QUIT

02.05 IF CMAX-N0CJ+3*K)> 2.10J RETURN

02.10 SET MAX=NOCJ+3*K>» RETURN

*GO

ENTER 3 ROWS AND 4 COLUMNS OF NUMBERS.

!0 s5 :8 s9
il :2 S3 !4
!9 !8 :7 16

LARGEST NUMBER IS = 9.00000*

*GO

ENTER 3 ROWS AND 4 COLUMNS OF NUMBERS.

SA SB :C sD

sA SB !C sD

SA SB sC SD

LARGEST NUMBER IS = 4.00000*

*GO

ENTER 3 ROWS AND 4 COLUMNS OF NUMBERS.

SA SB sC sD
84 S3 S2 51
SA 84 !C 52
LARGEST NUMBER IS = 4.00000*

9-46

Example 3:

C-FOCAL , 8/68 .

01.02 TYPE P'ROUTINE TO SOLVE MATRIX EQ. AX=B FOR X"!

01.04 ASK "ENTER DIMENSION OF A, THEN

01.05 TYPE ! "ENTER COEFF'S AC J,K> . . . AC J, N> AND BCJ>"!

01.10 ASK L.U SET N = L-1J SET I=-l

01.11 FOR K=0,Nj SET RCK)=K+1

01.12 FOR J=0,Nj TYPE !j FOR K=0,L; ASK ACJ+L*K>
01.14 SET M=lE-6

01.16 FOR J=0,Nj FOR K=0,N* DO 4.0

01.17 SET RCP)=0

01.18 FOR K=0,L; SET ACP+L*K> =ACP+L*K) /M
01.20 FOR J=0,N; DO 5.0

01 .22 SET 1=1 + 1

01.23 IF CI-N) 1.14,1.26,1.14

01.26 FOR J=0,N; FOR K=0,NJ DO 7.0

01.28 FOR K=0,N; TYPE ! %2, "X C "K, "> ",%8.05,XCK>

01 .29 TYPE ! ! i QUIT

04.0*5 IF CRCJ]) 0,4.3,4.1

04.10 IF <FABS<ACJ+L*K]>-FABSLM]) 4.3J

04.20 SET M=ACJ+L*K>

04.22 SET P=Jj SET Q = K

04.30 RETURN

05.10 IF CJ-P) 5.2,5.4,5.2

05.20 SET D=ACJ+L*Q>

05.30 FOR K=0,L; SET AC J-t-L*K> =AC J+L*K) -ACP+L*K ) *D

05.40 RETURN

07.10 IF C lE-6-FABS[A<J+L*K>]> 7.2* RETURN
07.20 SET XCK)=AC J+L*L)

*GO

ROUTINE TO SOLVE MATRIX EQ . AX=B FOR X

ENTER DIMENSION OF A, THEN

ENTER COEFF'S AC J,K> . . . AC J, N) AND BCJ)

1 :2 : 3 :4

4 :3 :2 : 1

1 :4 :3 :2

XC= 0) = 0.00000

XC= 1) =- 1.00000

XC= 2> = 2.00000

9-47

Demonstration Dice Game

Sooner or later, people who have access to a computer will try to
"match brains" with it or use it for their own enjoyment. Such pastimes
are usually keyboard oriented and FOCAL lends itself nicely to these
ends. The following example uses the random number generator,
FRAN( ), to produce dice combinations, plus IF logic to check bets
and winning combinations.

Lines beginning with a C indicate that the line is to be treated as a
comment and is not to be interpreted or executed. If a comment state-
ment is preceded by a statement number, the line is stored as part of
the program but does not affect the program logic.

The random number generator must be modified for use with statis-
tical or simulation programs to achieve true randomness. However, it
is sufficiently random in its present form for most applications.

NOTE

We naturally cannot assume any responsibility for the use of this or any

similar routines.

C-FOCAL . 8/68

01.10 SET B=0»TYPE !1"D1CE GAME"! ."HOUSE LIMIT OF $1000

01.13 TYPE ". MINIMUM BET IS $1"!1

01.20 ASK "YOUR BET IS"AJIF C1000-A) 3.1

01.22 IF <A-1>3.4.1 .26.1 .26

01.26 IF (A-FITR<A>>3. 5,1 .3.3.5

01.30 ASK MJDO 2JSET D=C»DO 2ITYPE " "ISET 0=D+C

01.32 IF {0-751.42.3.2.1.42

01.40 IF CD-2) 1 .5.3.3.1 .5

01.42 IF CD-I 1)1 .4.3.2.1 .4

01.50 IF CD-3) 1.6.3.3.1.6

01.60 ASK MJDO 2JSET E=CIDO 2JTYPE " "ISET E=E+C

01.72 IF CE-7) 1.74.3.3.1.74

01.74 IF CE-DJ1 .6.3.2.1 .6

02.10 SET C=FITRC10*FABSCFRANO>>JIF CC-6) 2 . 2. 2.2.2 . 1
02.20 IF <C-1)2.UTYPE XI." "ClRETURN

03.10 TYPE "HOUSE LIMITS ARE S1000"!l» GOTO 1-2

03.20 S B=B+A»T X7.1-Y0U WIN. YOUR WINNINGS ARE "B. ! ! J GOTO 1.2

03.30 S B=B-AJT J7,I"S0RRY. YOU LOSE. YOUR WINNINGS ARE "B. I 1 1 GOTO 1.2
03.40 TYPE "MINIMUN BET IS Sl'MUGOTO 1.2
'03.50 TYPE "NO PENNIES. PLEASE"! I J GOTO 1-2
*

*C ONCE YOU PLACE A BET. FOLLOW THE OTHER COLONS WI TH A
*C CARRIAGE RETURN TO INDICATE THE COMMAND "ROLL THE DICE".

9-48

*G0

DICE GAME

HOUSE LIMIT OF $1000. MINIMUM BET IS $1

YOUR BET IS:. 50 MINIMUN BET IS $1

YOUR BET IS: 15 :

= 6=3:

= 1 = A :

= 4=5
YOU WIN. YOUR WINNINGS ARE = 15

YOUR BET IS: 5 :

= 2=2:

= 6=1
SORRY, YOU LOSE. YOUR WINNINGS ARE = 10.

YOUR BET IS:3 : • ' .

= 6=5
YOU WIN. YOUR WINNINGS ARE = 13

YOUR BET IS: I'LL QUIT WHILE I'M AHEAD. THANKS!

9-49

.01

.02

.03

.08

.09

• 11

.70

.75

• 80

.90

.93

Schroedinger Equation Solver

The program is designed to aid the user in searching for possible
energy-states of an electron in a potential well. This is one of the most
complex equations yet written in FOCAL. It calculates and plots the
energy levels of an electron within specified boundary conditions.

C-FOCAL , 8/68

! j "SCHROEDINGER EQUATION SOLVER -", !

!," tDELSQUARED PSI + AX * PSI = E * PSI",!!

"TILTED SQUARE WELL PROBLEM WITH WIDTH", X0,!
"WELL TILT SLOPE A", Al , ! , "TRI AL ENERGY E",B1,!
"NUMBER OF STEPS", NT, !
VF=0J S SL=1

P<0)=0J S DX=X0/NTJ S PC1)=SL*DX> S R0=0
VF=0
P0=0

N=0, 1.NT-2J D 6
!,"PSI ZEROS"%2.0, P0
01.95 SOTO 7.02

05.10 T !,%3.0, PX," PSI",X,P(PX>,"."

05.20 S PZ=FITRCPM*SC)J S PE=FI TR<CP[PX ]+PM) *SC>

05.30 F X=l» 1,PZJ T " "

05.40 T ".",01 F X=l,l,PE+24J T " "

05.50 T "*"J R

06.10 S P<N+2>=<C-B1+A1*DX*CN+1 J>*DX«2+2>*PC N+l }-P(N)

06.20 I (NT-N-2) 12.90,6.9,6.3

06.30 S RB=PCN+2)*PCN+1 )J I (RB) 6.4,6.4,6.9

06.40 S P0=P0 + U R

06.90 CONTINUE

07.02 S CF = <P<NT>/P<1>> t2J T " CONV IND"%, CF

07.05 A " NEW E?"NY

07.07 I CNY-9) 7.9,7.08,7.9

07.08 I CVF) 7.09,7.8.7.09

07.09 I CCF-100) 7.1,7.1,7.8

07.10 S R2=PCNT)*VFJ I <R2) 7.73,7.80,7.85
07.73 S DB=-0.5*DB5 GOTO 7.85

7.80 S 08=0.1

07.85 S Bl =B1*C1+DB); T BIJ S VF(NT)J G 1.80

07.90 DO 14J GOTO 12.01

12.01 T !,!,"EIGEN E"B1I S HP=B1 /C Al *X0 )
12.20 T " EN/MAX P0T"HP, !
12.90 QUIT

14.10 S PM=0J S PP=0J F PX=l,l,NTj D 15

14.20 S PS=PM+PP* S SC=45/PS

14.30 T I!!J F PX=1,1,70J T "."

14.40 F PX=0,1,NT1 D 5

14.50 T UF PX = 1, 1,701 T "."

14.60 T !! J R

15.10 I (PCPX3) 15.2,15.9,15.5

15.20 I CPM+P<PX>> 15.3,15.4,15.4

15.30 S PM=FABSCPCPX3)

15.40 RETURN

15.50 I <P<PX>-PP> 15.9,15.9,15.6

15.60 S PP=PCPX)

15.90 RETURN

* 9-50

*G0

SCHROEDINGER EQUATION SOLVER -

-DELSQUARED PSI + AX *' PSI = E * PSI
TILTED SQUARE WELL PROBLEM WITH WIDTH.il
WELL TILT SLOPE A:50
TRIAL ENERGY E ! 50
NUMBER OF STEPS: 24

PSI ZEROS= 1 CONV IND= 0.329830E+02 NEW E?sY

= PSI*

0.000000E+00- *

1 PSI =

0.416667E-01 .

= 2 PSI =

0.798671E-01 .

3 PSI =

0.111713E+00.

= 4 PSI =

0-.135073E+00.

= 5 PSI=

0.148662E+00.

6 PSI =

0.152035E+00.

= 7 PSI =

0.145510E+00.

B PSl- =

0.130037E+00.

= 9 PSI=

0.10 70 40E+00.

= 10 PSI =

0.782351E-01 .

= 11 PSI=

0.454687E-01 .

= 12 PSI =

0.105644E-01 .

= 13 PSI=

-0.247985E-01 • * •

= 14 PSI =

-0.591747E-01 • ♦

= 15 PSI=

-0-914107E-01 . *

= 16 PSI=

-0.120671E+00. *

= 17 PSI=

-0.146440E+00. *

= 18 PSI =

-0.168S01E+00. *

= 19 PSI=

-0.186905E+00. *

= 20 PSI =

-0.201929E+00. *

= 21 PSI =

-0.214032E+00. *

= 22 PSI=

-0.223812E+00. *

= 23 PSI =

-0.231973E+00.*

= 24 PSI =

-0.239295E+00.*

EIGEN E= 0.500000E+02 EN/MAX POT= 0.100000E+01

9-51

SUMMARY OF COMMANDS, OPERATIONS, AND FUNCTIONS

Command

Example of Form

EXPLANATION

ASK

ASK X.Y.Z

COMMENT

CONTINUE

DO 4.1

ERASE

ERASE 2.0

ERASE 2.1

ERASE ALL

FOR i=x,y,z;(commands)

FOR i=%j,;( commands)

FOCAL types a colon for each variable;
the user types a value to define each vari-
able.

If a line begins with the letter C, the re-
mainder of the line will be ignored.

Dummy line?

Execute line 4.1; return to command fol-
lowing DO command,

Execute ail group 4 lines; return to com-
mand following DO command, or when a
RETURN is encountered.

Erases the symbol table.

Erases all group 2 lines.

Deletes line 2.1.

Deletes all user input.

Where the command following is executed
at each new value.

x=initial value of i

y=value added to i until i is greater than

GO
GO?

GOTO
IF

GO
GO?

IF(X)Ln,Ln, Ln

IF <X)Ln, Ln; {commands)

IF (X)Ln; (commands)

MODIFY

MODIFY 1.15

QUIT

RETURN

SET

SET A=5/B'C;

TYPE

TYPE A+B-C;
TYPEA-B.C/E;
TYPE "TEXT STRING'

WRITE

WRITE
WRITE ALL

WRITE 1.0

WRITE 1.)

9-52

Starts indirect program at lowest num-
bered line number.

Starts at lowest numbered line number
and traces entire indirect program until
another ? js encountered, until an error
is encountered, or until completion of
program.

Starts indirect program (transfers control
to line 3.4), Must have argument.

Where X is a defined identifier, a value,
or an expression, followed by three line
numbers.

If X is less than zero, control is trans-
ferred to the first line number.

If X is equal to zero, control is to the
second tine number.

If X Is greater than zero, control is to the
third line number.

Enables editing of any character on line
1.15 (see below).

Returns control to the user.

Terminates DO subroutines, returning to
the original sequence.

Defines identifiers in the symbol table.

Evaluates expression and types out = and
result in current output format.

Computes and types each expression sep-
arated by commas.

Types text. May be followed by ! to gen-
erate carriage return-line feed, or # to
generate carriage return.

FOCAL types out the entire indirect pro-
gram.

FOCAL types out all group 1 lines.

FOCAL types out line I.I.

FOCAL Operations

To set output TYPE %x.y where x is the total number

format, of digits, and y is the num-

ber of digits to the right of
the decimal point.
TYPE %6.3, 123.456 FOCAL types: = 123.456
TYPE % Resets output format to float-

ing point.
To type sym- TYPE $ Other statements may not fol-

bol table, low on this line

To input from high-speed paper tape reader,

** The second * was typed by the user. Input is from

the high-speed reader until occurrence of next *.
FOCAL types * for each line number read in from
the reader.

*1.10*; User typed 1.10*;. Input is taken from the high-

speed reader until occurrence of next *.

Modify Operations

After a MODIFY command, the user types a search character, and
FOCAL types out the contents of that line until the search character is
typed. The user may then perform any of the following optional
operations.

1. Type in new characters. FOCAL will add these to the line at
point of insertion.

2. Type a CTRL/L. FOCAL will proceed to the next occurrence
of the search character.

3. Type a CTRL/BELL. After this, the user may change the
search character.

4. Type RUBOUT. This deletes characters to the left, one chai-
acter for each time the user strikes the RUBOUT key.

5. Type «- . Deletes the line over to the left margin, but not the
line number.

6. Type RETURN. Terminates the line, deleting characters over
to the right margin.

7. Type LINE FEED. Saves the remainder of the line from the
point at which LINE FEED is typed over to the right margin.

9-53

The Trace Feature

Special
Character

Example
of Form

?...?

or
?...

Explanation

Those parts of the program enclosed in question
marks will be printed out as they are executed.
If only one ? is inserted, the trace feature becomes
operative, and the program is printed out from that
point until another ? is encountered, until an error
is encountered, or until program completion.

Summary of Functions

Square Root

Absolute Value
Sign Part

Integer Part

Random Number
Generator
Exponential
Function (e x )
Sine

Cosine

Arc Tangent

Logarithm
Analog-to-Digital

FSQT(x) where x is a positive number or ex-
pression greater than zero.

FABS(x) FOCAL ignores the sign of x.

FSGN(x) FOCAL evaluates the sign part only,
with 1.0000 as' integer,

FITR(x). FOCAL operates on the integer part

of x, ignoring any fractional part.

FRAN( ) FOCAL generates a random number.

FEXP(x) FOCAL generates e to the power x.
(2.71828")

FSIN(x) FOCAL generates the sine of x in

radians.

FCOS(x) FOCAL generates the cosine of x in

radians.

FATN(x) FOCAL generates the arc tangent of
x in radians.

FLOG(x) FOCAL generates log e (x) .

FADC(n) FOCAL reads from an analog-to-
digital channel, the value of the
function is that integer reading.

9-54

Scope Functions

FDIS(y) Displays y coordinate od scope and

intensifies x-y point.
FDXS(x) Displays x coordinate on scope.

Other functions are available from DEC for use with such periph-
erals as incremental plotter, card reader, etc.

Special Characters

1. Mathematical operators:
f Exponentiation

* Multiplication
/ Division

+ Addition

— Subtraction

2. Control characters:

% Output format delimiter

! Carriage return and line feed

* Carriage return

$ Type symbol table contents

( ) Parentheses ~\

[ ] Square brackets I (mathematics)

< > Angle brackets J

" " Quotation marks (text string)

? ? Question marks (trace feature)

* Asterisk(s) (high-speed reader input)

3. Terminators:

SPACE key (names) ]

RETURN key (lines) V (nonprinting)

ALTMODE key (with ASK statement) J

, Comma (expressions)

; Semicolon (commands and statements)

9-55

Error Diagnostics

Error messages are typed in the following format:

?nn.nn @

nn.rm (error code @ line number)

Error Code

Meaning

?00.00

Manual start from console

?01.00

Interrupt from keyboard via CTRL/C

?01.35

Group zero is an illegal line number

?01 .43

Illegal step or line number

?01.89

GOTO not used as one word or bad argument in IF

?01.;2

, Line number too large

?01.;3

Double periods in line number

702.48

Nonexistent line referenced by DO

702.63

Nonexistent group referenced by DO

702.81

Storage filled by push-down list

703.09

Nonexistent line used or a tight loop

703.31

Illegal command used

704.07

No space after IF or illegal format

704.35

Left of = in error for FOR or SET

704.48

Excess right parenthesis

704.56

Illegal terminator in FOR

705.63

Bad argument to MODIFY

706.13

Illegal use of function or number

706.64

Storage filled by variables

707.14

Operator missing or double E

707.34

No operator before parenthesis

707. <0

Double operators

707.; I

No argument given after function call

707. ;8

Illegal function name

708.50

Parentheses do not match

709.16

Bad argument in ERASE

709.50

Maximum group number exceeded

711.20

Input buffer has overflowed

712.83

Storage filled by text

720.41

Logarithm of zero requested

723.35

Literal number is too large

726.91

Negative exponent used

726.96

Exponent is too large

728.58

Division by zero requested

730.48

Imaginary square roots required

?31.<7

Illegal character or unavailable command or

unavailable function used

NOTE:

The above diagnostics apply only to the version of FOCAL, 8/68, issued on

tape DEC-08-AJAC-PB.

9-56

ESTIMATING THE LENGTH OF USER'S PROGRAM

FOCAL requires five words for each identifier stored in the symbol
table, and one word for each two characters of stored program. This
may be calculated by

5s +— • 1 .01 = length of user's program

where s = Number of identifiers defined

c = Number of characters in indirect program

If the total program area or symbol table area becomes too large,
FOCAL types an error message.

FOCAL occupies core locations l-3200 8 and 4600 8 -7576 8 . This
leaves approximately 700i locations for the user's program (indirect
program, identifiers, and push-down list). The extended functions
occupy locations 4600-5377. If the user decides not to retain the ex-
tended functions at load-time, there will be space left for approximately
1 100 10 characters for the user's program.

The following routine allows the user to find out how many core
locations are left for his use.

♦FOR 1=1,300; SET A(I) =1

?06.64 (disregard error code)

•TYPE % 4, 1*5, " LOCATIONS LEFT "

= 720 LOCATIONS LEFT *

9-57

CALCULATING TRIGONOMETRIC FUNCTIONS IN FOCAL

Function

FOCAL Representation

Argument
Range

Function
Range

Sine

FS1N(A)

0<|A|<10f4

0<|F|<l

Cosine

FCOS(A)

0<|A|<IOf4

0<|F|<l

Tangent

FSIN(A)/FCOS(A>

0<|A|<10t4
|Aj=(2N+l)r/2

0<|F[<IOf»

Secant

l/FCOS(A)

0<|A|<10t4

jA|=(2N+l)»/2

l<lF|<IOf6

Cosecant

1/FSINtA)

0<[A|<10t4
|A|=2N»

K|F|<10f6

Cotangent

FCOS(A)/FSIN<A)

0<|A|<10t4

|A| = 2J*r

0<|F|<!0I440

Arc sine

FATN(A/FSQT(I-At2)

0<|A|<1

0<|F|<i/2

Arc cosine

FATN(FSQT(l-At2)7A)

0<|A|<l

0<jF|<»/2

Are tangent

FATN(A)

0<A<10f6

0<F<»/2

Arc secant

FATN(FS0T(At2-D)

KA<I0t6

0<F<*/2

Arc cosecant

FATN(I/FSQT(AT2-I>)

KA<IOf300

0<F<»/2

Arc cotangent

FATNO/A)

0<A<10t615

0<F<»/2

Hyperbolic sine

(FEXP(A)— FEXP(- A))/ 2

0<|A|<700

0<|Fl<5»10t300

Hyperbolic cosine

(FEXP(A)+FEXP<- A))/ 2

0<|A|<70O

KF<J*10t300

Hyperbolic tangent

(FEXP(A)-FEXP(-A))/
(FEXP(A)+FEXP(-A))

0<|A|<700

0<|F|<1

Hyperbolic secant

2/<FEXP(A)-rFEXP(~A))

0<|A|<700

0<F<1

Hyperbolic cosecant

2/(FEXP(A)-FEXP(-A))

0<|A|<700

0<|F|<10t7

Hyperbolic cotangent

(FEXP(A)+FEXP(-A) )/
(FEXP(A)-FEXP(-A!)

0<jA|<700

K|F|<10t7

Arc hyperbolic sine

FLOG(A+FSQT(At2+l ) )

-10t5<A<IOt600

-I2<F<I300

Arc hyperbolic cosine

FLOO( A+FSQT( Af2- 1 ) )

KA<I0t300

0<F<700

Arc hyperbolic tangent

(FLOG(l+A)-FLOO(l-A))/2

0<|A|<1

0<jF|<8.31777

Arc hyperbolic secant

FLOG((l/A) + FSQT((l/At2»-l))

0<|A|<l

0<F<700

Arc hyperbolic cosecant

FLOG(( 1/A)+FSQT(( I/Af2)+I ))

0<|A|<IOf300

0<|F|<I400

Arc hyperbolic cotangent

(FLOO(X+l)-FLOG(X-l))/2

KA<10t616

0<F<8

9-58

Chapter 10

PDP-8 Computers
in the Sciences

As a leader in the field of small, general purpose computers, Digital
is well aware of the significant impact of small computers on the sci-
ences, particularly in the way scientists work in their laboratories and
in the way engineers design and build instruments, machines, and con-
trol systems. Scientists use PDP-8 family computers as personal, pow-
erful tools; engineers build them in as components. More computers in
the PDP-8 family have been sold for these applications than any other
computer in the world. Designed particularly for online applications;
PDP-8 computers embody design features that make them simple and
straightforward to interface and program for any application.

OFFLINE AND ONLINE USES

Before we discuss in detail the scientific applications of computers in
the PDP-8 family, let us define two basic concepts in this regard, off-
line and online.

A computer used to analyze data that has previously been recorded
is said to be used offline. In contrast, a computer used by a scientist to
directly collect, sample, and analyze data while an experiment is in
process is said to be an online computer.

There are two types of online usage. In the first type, the computer—
usually a small model — is physically located in the laboratory. In the
second type, where the amount of data to be processed is too large for
a small computer to handle, the scientist connects his experiment to a
larger, remotely located computer; this larger computer is usually
capable of handling several users simultaneously, i.e., on a time-shared
basis. A variation of this second type of online usage is to connect a

10-1

small computer situated in the laboratory to a larger, time-shared com-
puter at a remote location.

Offline use of a computer reduces computation time as compared to
manual calculations, but offers few other advantages. On the other
hand, an online computer permits the scientist to participate to a
greater degree in his experiment. The results which he constantly re-
ceives during the course of his experiment allow him to make certain
choices on the spot. He may call for a more intensive investigation of
interesting results, extend the scope of his experiment on the basis
of developing information,, or even instruct the computer to look for
and investigate interesting phenomena automatically during the experi-
ment.

DATA COLLECTION

Frequently, an experiment involves the measuring of a "signal" that
is continuous, although it may vary in amplitude, frequency, or both.
This signal must be converted into an electronic signal before it can be
processed by any type of computer, whether analog or digital; this con-
version is usually performed by a transducer or a potentiometer, and,
if necessary, the signal is then amplified by a high-gain amplifier. In
the case of digital computers, the analog signal thus produced must be
further converted by connecting it to the input channel of an analog-
to-digital (A/D) converter to digitize the signal. The A/D converter
changes a signal varying continuously with time to a series of discrete
numeric values. This conversion, known as "sampling a signal," can be
performed at many points in time during the experiment.

In addition to converted analog inputs as described above, a digital
computer can accept digital or contact-closure information. Process-
related on/off signals such as alarms, limit indications, and selector
switch settings can then be correlated with the analog data. Direct
digital transducers (e.g., digital shaft position encoders) can interface
to digital input channels without the intermediate A/D conversion.

Once the inputs are brought into the computer, the data can be veri-
fied, manipulated, formatted, and displayed in a variety of ways. The
arbitrary voltage or frequency readings from the A/D converter can be
compared against limits and converted to meaningful units (e.g., 3.26
millivolts from one sensor may correspond to +132°F; the same volt-
age from another sensor might represent 84 gallons per minute).
Through programming, input data can be compared, sorted, arranged
in labeled columns along with the time of reading, and displayed on a

10-2

Teletype printer or cathode ray tube. Data can be accumulated and
then written out on some peripheral medium for later analysis Qffline.
Communications linkages allow data to be sent to remote processors
as it is collected.

DATA ANALYSIS '

The computer can be used to analyze collected data to obtain arith-
metic means, medians, variances, deviations, correlations, regressions,
factors, etc., between one group of responses and another. Large sta-
tistical evaluations can be made either online or offline, depending
upon their complexity. A set of data can be compared with a pre-
viously collected data set, a user determined standard, or a component
within the sample itself. The user can specify the number of places of
accuracy he requires and have all calculations performed to this limit.

DATA DISPLAY

Using an online computer, the experimenter can display his data
while the experiment is running. Data can be displayed in a meaningful
format such as columnar listings on the line printer or as graphs on
an incremental plotter or cathode ray tube oscilloscope.

COMPUTER INSTRUMENTATION CONTROL

To illustrate the automation of instruments in the lab by a computer,
the following elemental example of data collection, analysis, and in-
strument control is given. The application consists of controlling the
pressure of a liquid within a pipeline, as shown in Figure 10-1. There
are four valves available for controlling the flow. A pressure gauge
transmits the current pressure. Assuming that the pressure is to be
always kept below 20 lb/sq.in., the flow chart in Figure 10-1 shows
the processing involved.

LABORATORY AUTOMATION

In laboratory work, the computer offers many advantages besides
just those of faster numeric computation. It can be programmed to
make judgments based on data as it is received. For example, it can
skim areas of little interest and concentrate on more promising areas.
The results of a general scan can be viewed by the experimenter while
a more detailed scan of some particular area is being performed. Also,
procedures can be modified or even terminated prematurely as the
progress of the experiment dictates.

10-3

COMPUTER

<"N

PRESSURE
GAUGE

^ >

i <

iv 1

iv 2

PIPELINE SCHEMATIC

SET

INITIAL

OPENING VALUES

STEP TO NEXT
DESIGNATED VALVE

IN SEQUENCE
(E.S.1,a,3,4,1.2)

INCREASE

OPENING VALUE

BY SET

AMOUNT

OPEN DESIGNATED

VALVE TO

OPENING VALUE

RETAIN
VALUES

RING ALARM

PRESSURE CONTROL
FLOW CHART

Figure 10-1. Example of Computer Instrumentation Control

10-4

Such flexibility is not possible with fixed-program devices; these
devices cannot vary their operations even though meaningful data is
not being collected. An instrument under the control of a computer, on
the other hand, can be made to find the proper areas of study. In addi-
tion, a readout device may be used to inform the experimenter of what
is happening and signal him immediately of any unusual results.

Computerizing instrumentation is a major step forward in laboratory
automation, but the loop between input and results can be closed even
further. For example, in the clinical laboratory, personnel shortages
and increasing hospital populations have placed an unprecedented
burden on the pathologist. This burden has been relieved by a clinical
computer, which can receive testing instructions, start and stop instru-
mentation, analyze the data, log the data in a precise format, correlate
the data with other physiological information, send the correct informa-
tion to a ward, coordinate with other departments, and keep accurate
and complete records.

The PDP-8 family of computers has become increasingly important
in this field of analytical instrumentation in the laboratory. Currently,
PDP-8 computers are being used in a variety of applications because
of their ability to process large numbers of chemical analyses repeti-
tively on samples with a standard matrix and a relatively narrow varia-
tion between samples. A sampling of these applications is given below,

Gas-Liquid Chromatography

The GLC-8 computer-based system can simultaneously handle 20
laboratory gas or liquid chromatographs. The system automatically
detects peaks and shoulders, calculates peak areas and peak retention
times, allocates peak overlap areas, corrects for baseline shift and
applies response factors. Using stored analysis tables as internal
standards, peaks can be identified and the percentage of each com-
ponent determined. GLC-8 prints out a complete analysis report,
including retention time, peak area, component percentage and toler-
ance. Analysis time is saved, human errors reduced, and efficiency of
instrument use greatly increased. Types and sampling rates of chromato-
graphs can be mixed, analysis tables can be calibrated and updated
automatically, and instantaneous analysis reports to remote locations
can be provided. GLC-8 is designed primarily for use with gas or
liquid chromatographs, but may be used with other analytical instru-
ments, such as amino-acid analyzers, that use the same method of
analysis.

10-5

Pulse Height Analysis

PHA-8 is a complete computer system for single- or multi-parameter
analysis. It can gather, store, display, and analyze energy or time-of-

PDP
COMPUTER

PAPER
TAPE
PUNCH

PAPER

TAPE
PUNCH

KSR-33
KEYBOARD
PRINTER

CRT
DISPLAY

TCOZ
DISPLAY

NUCLEAR
INTERFACE

TUS9
DECTAPE

DUAL 4096

CHANNEL

NUCLEAR ADC

1 DETECTOR

| DETECTOR |

TU5S
DECTAPE

I/O BUS

Figure 10-2. Computer-Based Pulse Height Analysis System

flight spectra and record the results on a variety of output devices.
Using analog-to-digital converters, PHA-8 can be configured with
4,096 to 32,768 channels for spectra storage. With optional mass
storage devices, systems can be set up to run multiple experiments auto-
matically, store results from each experiment, and retrieve them upon
command. Background counts can be subtracted, spectra compared,
one spectrum subtracted from another, and energy calculations per-
formed. Oscilloscope subsystems may be used to generate contours,
isometrics, and area-of-interest displays. The computer can also be
used to rotate axes, integrate areas between markers and under peaks,
calibrate energy versus channel number, and fit curves.

10-6

Mass Spectroscopy

The extremely high data rates required in high resolution mass
spectroscopy are easily handled by PDP-8 systems. With sampling
rates above 20KHz, the computer can scan 500 or more peaks with
from 20 to 50 points per peak. The computer can also be used to con-
trol the sweep of the magnet, providing lower data rates and increased
accuracy when desired. Centroid calculations, to determine peak posi-
tions accurately, may be performed while the scan is taking place.

In low and medium resolution mass spectroscopy, where the analyst
has much slower data rates, the computer can generate a mass map to
determine integer mass units. The computer can be connected to mag-
netic deflection, time-of-flight (TOF), or quadropole mass spectrom-
eters to observe ion current versus voltage potential.

MASS
SPECTROSCOPE

BULK
STORASE

MULTIPLEXER

•

POP
COMPUTER

r-l TAPE 1

TELETYPE

A/6
CONVERTER

CLOCK

Figure 10-3. Computer-Based System for Mass Spectroscopy

Nuclear Magnetic Resonance (NMR) Spectroscopy

Adding a PDP-8 computer-based system to an NMR spectrometer
significantly benefits the analyst in the collection, analysis, and presen-
tation of the data. In the data collection phase, the computer system
provides for more capability than a hard-wired signal averager, allow-
ing online selection and examination of key areas of interest and
improved sensitivity. In the analysis phase, corrections may be applied,
and Fourier analyses performed directly without prior storage or read-
out of the data. The computer can also be used to reduce the spectra
to compact form so that they may be stored, retrieved, and compared
with other data. Spin simulation may be calculated on relatively
complex spectra.

10-7

Absorption Spectroscopy

PDP-8 computer-based systems decrease noise by mathematical
smoothing, locate the position of maximum absorption, and auto-
matically record the absorption value. With the computer, the analyst
can easily apply background and frequency corrections: 100% and
zero line, wave number, ordinate (percentage transmittance), slit
function, and tracking error. The computer permits the analyst to
determine integrated band densities, resolve poorly separated bands,
and synthesize hypothetical spectra — all while the sample is being
run. The computer also performs difference analyses (subtracting one
spectrum from another), automatically replots spectra, tabulates peak
data, and identifies peaks by searching a spectrum library in the
computer's memory.

| RATE METER |

CONTROL
PANEL

1 SCALEB h — 1

•) PRINT CONTROL U

T~T

DIGITAL PUNCHED

PRINTOUT PAPER

TAPE TAPE .

DFFRACTOMETER

X-RAY
6ENERAT0R

\WLTA6E AND

CURRENT

STABILIZATION

Figure 10-4. Computer-Based X-Ray Diffraction System

X-Ray Diffraction

A PDP-8 computer-based system enables the analyst to obtain more
reliable data by allowing him to correct errors caused by accidental
changes m experimental setups; it also corrects for occasional slippage
of the sample as it is held in the path of the X-ray. To do this, the
computer is directed to analyze the data as it is received and to com-

10-8

pare it with previously established standards stored in memory. New
measurements that deviate from the stored information generate new
input control signals used in positioning the crystal.

Time-of-Flight Analysis

PDP-8 systems are available for performing time-distribution studies
for three types of experimentation: velocity measurements to determine
energies, life-time measurements to determine identities of unstable
particles, and time-coincidence measurements to seek correlations be-
tween events occurring near each other in time. All systems provide
scalars that allow them to collect data at high neutron counting rates,
to process several events during each cycle of the accelerator, and to
eliminate cumbersome deadtime corrections. Time-measuring sections
are available for low-energy particle experiments with channel widths
down to 0.1 microsecond. For high-energy applications, there are total
systems that include commercially available time-measuring sections.

OTHER APPLICATIONS

The efficiency, accuracy and speed of any instrumental analysis can
be significantly increased by including a PDP-8 computer in the
experimental system to control the sequence of events and record and
analyze the accumulated data. A PDP-8 computer can be easily inter-
faced to a wide range of analytical instruments, including tensile testers,
electron microprobes, ultracentrifuges, microdensitometers, liquid scin-
tillation counters, and emission spectrographs. Extensive uses in the
physical, life, and natural sciences have included the following applica-
tions.

Field Typical Applications

PHYSICS High- and low-energy studies. Coupled to

flying spot scanners, nuclear detectors and
counters, mass spectrometers, bubble
chambers, X-ray diffractometers, and ac-
celerators.
NUCLEAR Used at a radiation lab online for readout,

REACTIONS display, and graph plotting. Can monitor

a dual system at a nuclear reactor site.

10-9

Field

GAMMA-RAY
SPECTROSCOPY

ELECTROCHEMISTRY

CRYSTALLOGRAPHY

NEUROLOGY

CARDIOLOGY

CLINICAL

MEDICINE

Typical Applications

Used at a testing station in conjunction
with an oscilloscope and light pen, receiv-
ing gamma-ray spectra from high resolu-
tion detectors and A/D converters. "■
Used to investigate electrochemical phe-
nomena at the electrode-electrolyte inter-
face and to determine the chemical com-
position of steel samples.
Sets individual crystal angles of diffracto-
meter, records X-ray reflection data, and
plots crystal structure on CRT.
Used to study and determine power spec-
tra of finger and hand tremors, inter-
actions of communities of neurons, time-
interval histograms from discharges in
cerebellar units, and the effect of electrical
stimulation upon activity in the human
ulnar nerve.

In brain mapping, single cell activity with-
in the brain of a monkey is recorded
and impedance measurement data from
the brain is digitized to map brain struc-
tures on a scope.

Used to average and store changes in
transmembrane voltage of cardiac muscle
cells after external electrical stimulation,
to study the hydrodynamics and trans-
mission characteristics of the mammalian
arterial system, to evaluate by Fourier
analysis complex pressure pulses in the
aorta of mammals, to measure micro-
circulation, to evaluate velocity profiles
and viscuous losses in blood vessel walls
during pulsatile flow, and to evaluate
pressure flow relations and propagation
of pressure pulses in arterial trees of a live
subject.

Used to perform routine data processing,
record handling, and statistical research,
to analyze ECG's to aid the brain surgeon

10-10

Field

IMAGE
PROCESSING

BEHAVORIAL
. SCIENCES

EARTH
SCIENCES

OCEANOGRAPHY

ATMOSPHERIC
SCIENCES

Typical Applications

in the operating room in locating the
cerebral area in which to work, to study
Parkinson's disease, and to aid in endo-
crine analysis.

Used tc? analyze photos of cells and chro-
mosomes, to estimate karyotypes, to per-
form density measurements on autoradio-
graphs,to correlate photos of vocal cord
action with the sounds made, and to per-
form photogrammetric analyses.
Used to study interresponse times and the
social behavior of primates, to study the
operant conditioning of pigeons, to con-
trol and monitor schedules of reinforce-
ment, to determine auditory, tactile, and a
visual responses in the human, and to
perform online psychological and se-
quence pattern tests.

Used in seismology and paleontology to
identify small teleseismic disturbances, to
correlate^multiband images from airborne
or spaceborne cameras, to combine mul-
tiple polarization radar images, to group
and separate fossils, to aid in undersea
drilling by scanning and analyzing elec-
trical signals from sensors located on the
rig floor, and to compare and correlate
oil-field brines by means of pattern anal-
ysis.

Used to process data on sea conditions,
to assist in navigation and ship control,
to perform oceanographic surveys, to de-
termine the biomass in the oceans, to
measure and correlate simultaneously the
total magnetic intensity of the earth's field,
the gravity, and the sea surface tempera-
ture at a specific location while making a
profile of the subbottom structure.
Used to analyze air polution, to aid in
micrometeorological studies, and to an-
alyze fluid mechanic interactions.

10-11

Field

GEOPHYSICS

SPACE
SCIENCES

Typical Applications

Used to investigate the ionosphere, to
process telemetry data from scientific in-
strumentation in sounding rockets and
balloons, and to aid in research of solar
phenomena.

Used to check out guidance and control
equipment for the United States' largest
rockets, to perform online testing of space
satellite hardware, to decode data from
orbiting satellites, to check out scientific
payloads, to develop a solar wind spec-
trometer, to process photographic and
other signals from lunar and planetary
space probes, to handle radio telescope
control systems and collect and reduce
radio astronomy data, and to automate
optical telescope operations.

AN EXAMPLE

To illustrate the adaptation of a PDP-8 computer to a scientific ap-
plication, we have selected as an example the collecting and preproc-
essing of physiological signals by a computer for subsequent analysis
by another computer. 1

A PDP-8 program was written to receive and preprocess as many
as eight physiological signals simultaneously from a monitored patient
(or individual signals from up to eight patients). Inputs are analog
signals, which are sampled 500 times/sec. and digitized. The program
performs a code recognition of each signal and temporarily stores this
and subsequent data on a drum until the number points required for
the analysis have been accumulated. Concurrently with other instruc-
tions, the data break facility allows short blocks of data to be written
on the drum or long blocks read back into core. The long-block data
(referred to as a "lead"), needed to perform the analysis, is relayed via

1 The information in this section was extracted from a paper in "DECUS Pro-
ceedings, Fall, 1967" submitted by Miss Maxine L. Paulsen of the Medical
Systems Development Laboratory, Washington, D.C.

10-12

an interface to the other computer, which transfers the data to magnetic
tape. The tape is then used as input to a third computer, which con-
solidates, analyzes, and interprets the signals for each patient. The
operations of all three computers are carried on simultaneously once
the first input tape is written.

The Preprocessing Hardware

The equipment used in this application is a DIGITAL Preprocessing

System F, which includes the following components.
A 12K PDP-8 Processor
One Type DM01 Data Channel Multiplexer
One Type RM08E Serial Magnetic Drum
One Type 139E General Purpose Multiplexer Control
One Type 138E General Purpose A/D Converter
One Programmable Real Time Clock

One Signal Input-Routing Network Package (includes 24 Type
A 103 Multiplexer Switches and 64 Amplifier Mounting Boards)
One High-Speed 2-Way Interface to the other computer
One Interrupt and Skip Logic feature with 8 Program Flags

Also included in the system, but not used in this particular application,
, were one Type TC01 DECtape Control Unit, two Type TU55 DEC-

tape Transports, and one PC01 High-Speed Paper Tape Reader and

Punch.

Input/ Output Specifications for the PDP-8 Program

The input to this program is any physiological signal that is pre-
ceded by 3 or 13 BCD digits represented by square waves. Data, can
be received on up to eight channels simultaneously. The data are digi-
tized to a precision of 10 bits while being sampled at the rate of 500
times/sec. The output from this program is the "long block" transferred
to the other computer via the high-speed interface. The other computer
inputs the "long block" data, writes it on magnetic tape, and then sig-
nals the PDP-8 that it is ready for another block. This long block has
two words for channel identification, 13 words for the recognized BCD
digits, and 2,032 data points from the signal. These data points are
used in the analysis of the signal by the other computer.

Timing and Storage Considerations

An input block size of 128 words was chosen; eight blocks this size
(allowing one for each channel) can be transferred from the core to
the drum faster than the next eight blocks can be brought in from
the A/D converter (it takes the same amount of time to transfer one
or eight blocks into core, because of the sampling rate). Two blocks,

10-13

128 words each, are reserved in core for each channel and are utilized
in a double-buffered manner so that data is being written out of one
block onto the drum while at the same time data is being read into the
other block from the A/D converter. The block size was chosen from
the timing table given in Table 10-1.

Associated with these blocks is a "Block Ready Table," which keeps
a record of all blocks to be written on the drum; because of the al-
ternating use of the blocks as described above, only one block for each
channel should appear in this table at any one point in time. The en-
tries recorded in the table are in the form xxOc, where xxOO is the
starting address in core of the block to be written, and c is the asso-
ciated channel and is used indirectly to determine the starting location
on the drum where the block is to be written. Once the block is writ-
ten, the entry in the table is replaced by zeros. Two pointers are asso-
ciated with this table to ensure that the first block read into core is
the first block written out on the drum. Pointer 2 gives the location in
the table where the next block ready is to be recorded; pointer 3 gives
the location in the table which contains the location in core of the next
block to be written on the drum. The pointers travel through the table,
with pointer 3 following pointer 2, until the end of the table is reached;
then the pointers circle back to the beginning of the table again (see
Figure 10-5).

For each channel, sixtsen 128-word blocks are accumulated on the
drum in consecutive locations to make up a "lead" to be read back
into core, and then transferred to the other computer. As in the pre-
ceding discussion, there is an associated table, called a "Lead Ready
Table," used to keep a record of all "leads" ready to be transferred.
This table has 64 entries, (the number of leads that can be stored on
the drum — a maximum of eight leads per channel). The entries in
this table are- of the form yyfO, where yyOO is the starting location on
the drum of the lead, and f is the field (f is set-to 4 for field to avoid
having an entry of 0000 when the lead starts at location 0000, field 0).
When a lead has been read into PDP-8 core from the drum and trans-
ferred to the other computer, that entry in the table is set to 0000.
As with the other table, two pointers are also associated with this table.
Pointer points to the location in the table where the next lead to be
completed is to be recorded; pointer 1 points to the location in the
table which contains the location on the drum of the next lead to be
read and transferred. This arrangement ensures that the leads are
transferred in the same order they are completed (see Figure 10-5).

10-14

Program Initialization

The program begins by clearing all peripheral equipment flags.
Next, the usual initialization, such as clearing tables, setting counters
and initial exits, etc., is done. Then, the drum flag is set by writing a
sector on the drum. The clock is set to interrupt 500 times/sec. The
program halts and, when everything is ready, the CONT (continue)
switch is pressed. The clock is started, the interrupt turned on, and the
program is sent to the write routine, where the first interrupt will occur
(see Figure 10-6).

Program Interrupt Service Routines

The interrupt is off during interrupt servicing.
Answer Interrupt Routine. The clock interrupt flag is cleared, the
contents of the accumulator are saved, and the return from interrupt
is set up (see Figure 10-7).

A/D Service Routines (one for each channel). A value from the signal
is digitized (0 through 10 bits) and stored, right adjusted. When all
channels have been serviced, the program goes on to test the eight
words just read in (see Figure 10-7).

Test Routines or Path Selector (one for each channel).

EXIT1 - "No data" path. Originally set to come to this exit. As soon

as a nonzero value is received on this channel,' EXIT2 is set

(see Figure 10-8).

EXIT2 - "Code recognition" path. While set to this path, the data
values are examined point by point, and the BCD digits rep-
resented by the square waves are recognized. After the re-
quired number of digits have been found and stored in the
first block for a "lead," EXIT3 is set (see Figure 10-8).

EXIT3 - "Data store" path. The first block of data has two words
for channel identification; the next 13 words are for the
BCD digits just recognized; the remaining 113 words are
filled with data points from the signal. When these 16 blocks
(a "lead") have been input, EXIT2 is set to wait for the
next BCD code. As each block is filled, it is recorded in
the "Block Ready Table," pointer 2 is incremented, and the
functions of Block 1 and Block 2 are switched (see Figure
10-8).

Exit from Interrupt Service. When each of the eight words is processed,
the interrupt is turned on, and the program returns to where it was
previously interrupted (see Figure 10-8).

10-15

Program I/O Routines

Drum Write. By using pointer 3, the program checks the "Block Ready
Table" to see if there are any blocks to write. If not, the program
goes to the Drum Read routine. If there are, the block is written from
the starting core location xxOO given in the table to a location on the
drum found indirectly by using the channel number c, which is also
found in the table entry. After writing, the entry in the table is set to

0000 and pointer 3 is incremented. The drum location for this channel
is appropriately advanced. When 16 blocks have been written, a "lead"
has been stored on the drum. It is recorded in the "Lead Ready Table"
in the location indicated by pointer 0; then pointer is incremented.
The program then goes back to the beginning of this routine to see if
there are any more blocks to write. See Figure 10-9.

Drum Read. By using pointer 1, the program checks the "Lead Ready
Table" to see if there are any leads to read and transfer. If not, the
program goes to the Drum Write routine. If a "lead" is ready, and the
last transfer is completed, the "lead" is read in from location and
field yyfO of the drum into the space reserved for a "lead" in core.
Then, this entry in the "Lead Ready Table" is set to 0000, and pointer

1 is incremented. See Figure 10-9.

Transfer to Other Computer. The program sets up and initiates the
transfer of the "lead" from the PDP-8 core to the core of the other
computer, then goes to the Drum Write routine. See Figure 10-9.

Table 10-1. Example Timing Table

Block Size

Drum Transfer Ti

Data Input Time

(# words)

Core Drum

A/D Core

(max. access time = 17.3 ms)

1 Channel

8 Channels

1-8 Channels

17.3 + 0.25 = 17.55 ms

140.4 ms

32 ms

17.3+ 1.00=18.3 ms

146.4 ms

128 ms

1281

17.3 4- 2.00 = 19.3 ms

154.4 ms

256 ms

256

17.3+ 4.00 = 21.3 ms

170.4 ms

512 ms

512

17.3 + 8.00 = 25.3 ms

202.4 ms

1024 ms

1024

17.3 + 16.00 = 33.3 ms

266.4 ms

2048 ms

iThe input size used in this application, because 8 blocks in the least
number of blocks that can be written on the drum in less time (154.4 ms)
than it takes to read them into core (256 ms).

10-16

Example of block
ready table

LOC CONTENTS

0042 0165

0043 0161

(Pointer 2 — where to

record next block

ready)

(Pointer 3 — the next bl

to be written on the

drum)

xxOc

xxOO =

starting loc.

in cofe of

block
c = channel

number (0-7)

BLOCK
READY
TABLE

Example of lead
ready table

LOC CONTENTS

0040 3006 (Pointer 0— where to
record next "lead"
ready)

0041 3002 (Pointer 1— the '
to be read and
transferred next)

lead"

I 3001 0000

3002 0040

3003 0010

3004 2210

3005 4040

3006 0000

I 3075 0000
J 3076 0000
L 3077 0000^

yyfO
yyOO =
starting loc
on drum of
"lead" (2048
words
f = drum field
0(=4) or 1

LEAD
READY
TABLE

Figure 10-5. Examples of "Block Ready" and "Lead Ready" Tables

©H

SET DRUM FLAG

SET CLOCK TO

INTERRUPT 500

TIMES/SEC.

IOF INITIALIZE
' (A't)

SET CLOCK
ION

Figure 10-6. Example of Initialization

10-17

|ntrptJ — *

CLEfR CLOCK

INTERRUPT FLAG

SAVE (AC1

SET RETURN

A/D SERVICE (0-7)

CONVERT AND /■ — %

STORE EIGHT _JtEST 1

WORDS ^\ c /

(10-BIT PRECISION) V_-/

Figure 10-7. Example of Interrupt Answering and A/D Service Routines

C=0,l,...7 A=l,2,3

RESTORE (AC)
ION

STORE THE
13-DIGIT CODE

5 _j in The first

^ INPUT BLOCK FOR
THIS "LEAD"
SET A- 3

CONTINUE TO
RECOGNIZE CODE
(3 OF! 13 DIGITS)

READY »Oc IN
"BLOCK READY
TABLE" ADVANCE

POINTER 2
SWITCH BLOCKS

Figure 10-8. Interrupt Service
10-18

RECORD yyfO IN
"LEAD READY
TABLE 1 ' ADVANCE

WRITE BLOCK ON

DRUM IN PROPER

LOCATION

ADVANCE DRUM
LOCATIONS

SET ENTRY= O

IN "BLOCK READY

TABLE" ADVANCE

POINTER 5

READ IN "LEAD"

(2048 WORDS)

SET ENTRY- O

IN "LEAD READY

TABLE" ADVANCE

POINTER t

SET UP FOR

TRANSFER AND

INTERRUPT OTHER

COMPUTER

Figure 10-9. I/O Routines (Interrupt is on)
10-19

10-20

Chapter II

Digital Equipment

Computer Users

Society

OBJECTIVES

Digital Equipment Computer Users Society (DECUS) was estab-
lished to advance the effective use of Digital Equipment Corporation's
computers and peripheral equipment. It is a voluntary, non-profit users
group supported by DEC, whose objectives are to:

advance the art of computation through mutual education and
interchange of ideas and information,

establish standards and provide channels to facilitate the free ex-
change of computer programs among members, and

provide feedback to the manufacturer on equipment and program-
ming needs.

The Society publishes a newsletter, DECUSCOPE, every two
months, sponsors technical symposia twice a year (Spring and Fall),
maintains a program library, and publishes proceedings of its symposia.

DECUS PROGRAM LIBRARY

The DECUS Program Library is one of the major functions of the
users group. It is maintained and operated separately from the DEC
library and contains programs contributed by users. Programs are avail-
able for the PDP-8/I,-8/L,-8/S, and -8. (Programs are also available
for the PDP-1, PDP-4/7/9, PDP-6/10, PDP-5, LINC, and LINC-8.)
The library contains many types of programs, such as executive
routines, editors, debuggers, special functions, games, maintenance, and
various other classes of programs.

11-1

Programs for the PDP-8 family and the PDP-5 are listed later in
this chapter together with abstracts of a few programs that are fre-
quently requested by users.

Another feature of the program library is "Programs Available from
Authors." This was initiated in order to allow users access to pro-
grams which are not fully documented or fully debugged but which
are working to a certain extent. These programs are announced period-
ically in DECUSCOPE with information as to where they may be ob-
tained. The Program Library Catalog also contains a section describing
these programs.

Forms and information for submitting programs to the library may
be obtained from the Executive Secretary. Any user may submit a pro-
gram which he feels will be of use to others. Specifications for pro-
grams to be submitted jire fairly simple. It is required, however, that
documentation and operating instructions be clear. An object and
symbolic tape of each program submitted is desirable.

Programs are available to all members on a request basis. Requests
for programs should be made on DECUS Library request forms and
directed to the DECUS Program Library.

Programs submitted to the library are reviewed by the Programming
Committee before being placed in the library. A review checklist is
sent out with each program for evaluation by the user. Noteworthy
comments and suggestions on DECUS programs are published in the
newsletter. Certification of programs is under the jurisdiction of the
Programming Committee.

The library presently contains 351 programs. In 1967, 8815 pro-
grams were issued to requestors, and 127 programs were submitted to
the library.

DECUSCOPE

DECUSCOPE is the Society's technical newsletter, published since
April, 1962. The aim of this informal news "scope" is to facilitate the
interchange of information. The majority of articles are contributed
by the users who are invited to submit ideas, programming notes,
letters, and application rotes for publication. DECUSCOPE is mailed
every two months to members and others interested in DEC's compu-
ters and DECUS. Circulation reached 3,200 copies per issue in April,
1968.

Forms for submitting material to DECUS (newsletter or library) are
available from the Executive Secretary.

11-2

ACTIVITIES

Two nation-wide symposia are held each year — one in the SPRING
and the other in the FALL. Regional seminars and workshops are also
held periodically. The proceedings and papers presented at the sym-
posia and seminars are published shortly after each meeting and are
sent automatically to meeting attendees and upon request to others.

DECUS sponsored the first workshop meeting of the Joint Users
Group of the Association for Computing Machinery in April, 1966,
and has actively participated in workshops held each year since. The
purpose of the Joint Users Group meetings is to establish means for
intercommunication among user groups.

DECUS is also a member of the Joint User Group Library Catalog
Project sponsored by JUG. This catalog contains lists of programs
available from several major user groups. Members of the participating
user groups will be eligible to request program documentation from
other groups through their Program Interchange Chairman, i.e., for
DECUS members, the DECUS Executive Secretary. Specific details
on this interchange program are available from the DECUS office.

DECUS encourages subgrouping of users with common interests.
Special interest groups, such as the following, have been formed.

European Users have formed a group electing a committee to for-
mally organize meetings annually in Europe. They have held three
meetings to date with proceedings being published for each meeting.

The Education Sub-Group, organized in the early months of 1968,
held their first technical session and sub-group workshop at the DECUS
Spring 1968 Symposium in Philadelphia. Enthusiasm ran high, and
chairwoman Mrs. Judith Edwards of Computer Instruction NET-
WORK is highly confident that they will be an extremely active and
productive group.

Users in the Biomedical field have expressed the desire to hold meet-
ings specific to their field. Sessions with papers presented in this area
were held at two DECUS meetings, and a separate Biomedical Seminar
was held in New York City in 1967. A formal sub-group with a chair-
man has not been formed at this time, but it is hoped that such a group
will be organized in the near future.

Information on how to join or formally organize a sub-group may
be obtained by contacting the DECUS Executive Secretary.

11-3 .

MEMBERSHIP

Membership in DEC US is voluntary and does not require the pay-
ment of dues. Members are invited to take an active interest in the
Society by contributing to the program library, to DECUSCOPE, and
■by participating in its meetings and symposia. There are two types of
membership in DECUS Installation Membership and Individual Mem-
bership.
Installation Membership

An organization which has purchased or has on order a computer
manufactured by Digital: Equipment Corporation is automatically eligi-
ble for installation membership in DECUS. Membership status is ac-
quired by submitting a written application to the Executive Secretary
for approval by the DECUS Board.

An organization may appoint one delegate for each DEC computer
owned. The delegate should be one who is immediately concerned with
the operation of the computer he represents and who is willing to take
an active part in DECUS activities. He is entitled to vote on all DECUS
policies and during the election of. officers. The delegate receives
DECUS literature automatically and library programs upon request.

A three-ring binder containing information and forms pertinent to
the users group is sent to each delegate upon acceptance into the So-
ciety. The binder provides a convenient means for maintaining and
updating DECUS literature. It contains such material as bylaws, news-
letters, library catalog, forms for submitting and requesting material,
indexes of newsletters and proceedings, etc.
Individual Membership

There are two classes for individual membership:

1 . Individuals desiring membership in DECUS who are employed
at an installation but are not appointed delegates.

2. Individuals who have a direct interest in DECUS or DEC
computers but are not employees of a DECUS installation
member.

An individual member is not entitled to vote on DECUS policies or
during elections. They receive on an automatic basis only the news-
letter and Program Library Catalog. They can, however, receive other
DECUS material on a request basis.

Written application indicating desire to join must be submitted to
the Executive Secretary for approval by the DECUS Board. There is
no limit to the number of individual members that may join from either
an installation or a non- installation.

11-4

MEMBERSHIP — APRIL 1968
Installation Delegates — 1136 Individual Members — 1238

EXECUTIVE BOARD, POLICIES AND ADMINISTRATION

The Society's policies are formulated by an Executive Board elected
by vote of Installation Member delegates.

The board consists of a President, Executive Secretary, Recording
Secretary, and Standing Committee Chairmen. In addition, a non- vot-
ing representative of Digital Equipment Corporation is a member of
the board. The DECUS president for the preceding year is also in-
cluded as a non-voting member of the Executive Board.

The Administrative office is located at Digital Equipment Corpora-
tion, Maynard, Massachusetts 01754, and all correspondence should
be directed to the attention of the DECUS Executive Secretary.

DECUS PROGRAM LIBRARY CATALOG

The DECUS Program Library Catalog contains lists and abstracts of
all programs available from the DECUS Library. Programs for all
DECs computers are included. The catalog is divided into three sec-
tions: category index, abstracts and numerical index, and "Programs
Available from Authors." The category index for PDP-5/8 users pro-
grams and selected program abstracts are included below. Each mem-
ber of DECUS receives a catalog automatically. Additional copies may
be requested. The catalog is updated periodically and new additions
to the library are published in DECUSCOPE.

PDP-5, PDP-8, -8/S, -8/1, -8/L Category Index

Category DECUS No. Title ~~~

Executive Routines, 5-13 PDP-5 Assembler (for use on IBM 7044/7094)

Assemblers and 5/8-18A,B,C, Binary Tape Disassembly Programs

Compilers 5/8-20 Remote Operator FORTRAN System

5/8-28a* PAL HI Modifications - Phoenix Assembler

5/8-45* PDP-5/8 Remote and Time-Stiared System

5/8-46a Utility Programs for the PDP-5 and PDP-8

8-59 PALDT— PAL Modified for DECtape (552 Control)

5/8-64 DECtape Programming System

8-67 PAL Modified for DECtape Input

8/8S-77 PDP-8 Dual Process System

8-82 Library System for 580 Magnetic Tape (Preliminary Version)

8-84 One-Pass PAL HI

8-91 MICRO* An On-Line Assembler

8-102 A LISP Interpreter for the PDP-8

8-110 Directory Print (DIREC)

8-115 bouble Precision Integer Interpretive Package

8-116 PDP-8/ Automatic Tape Control (Type 57 A) Library System

8-122 SNAP (Simplified Numerical Analysis Program)

8-123 UN1DEC Assembler

8-124* PDP-8 Assembler for IBM 360/67

8-125 PDP-8 Relocatable Assembler for IBM 360/67

_^ 6/8-12 PDP-8 Assembler for PDP-6

11-5

PDP-5, PDP-8, -8/S, -8/1, -8/L Categoiy Index (com.)

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Books

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "dec :: pdp8 :: handbooks :: IntroToProgramming1969"

See other formats